WEBVTT 1 00:00:08.480 --> 00:00:11.599 - Hello, and welcome to the Physics World weekly 2 00:00:11.599 --> 00:00:12.099 - podcast. 3 00:00:12.480 --> 00:00:13.859 - I'm Hamish Johnston. 4 00:00:14.544 --> 00:00:17.285 - In this episode, I'm joined by Antonia 5 00:00:17.585 --> 00:00:19.845 - Seymour, who is chief executive 6 00:00:20.304 --> 00:00:21.685 - of IOP Publishing. 7 00:00:22.545 --> 00:00:25.425 - A wholly owned subsidiary of the Institute of 8 00:00:25.425 --> 00:00:25.925 - Physics, 9 00:00:26.545 --> 00:00:29.364 - IOP Publishing produces more than 100 10 00:00:29.839 --> 00:00:30.899 - scholarly journals. 11 00:00:31.440 --> 00:00:35.119 - It also publishes Physics World, so Antonia is 12 00:00:35.119 --> 00:00:36.420 - a colleague of mine. 13 00:00:37.200 --> 00:00:40.560 - Antonia has more than thirty years experience in 14 00:00:40.560 --> 00:00:42.979 - a publishing industry that has evolved 15 00:00:43.359 --> 00:00:43.859 - significantly 16 00:00:44.320 --> 00:00:47.504 - from the days of paper journals to today's 17 00:00:47.804 --> 00:00:48.304 - sophisticated 18 00:00:48.684 --> 00:00:49.744 - digital technologies. 19 00:00:50.684 --> 00:00:54.545 - In this interview, we explore how scholarly publishers 20 00:00:54.604 --> 00:00:57.984 - are responding to the rapid growth in research 21 00:00:58.125 --> 00:00:58.625 - output, 22 00:00:59.250 --> 00:01:01.429 - the rise of artificial intelligence, 23 00:01:01.969 --> 00:01:05.349 - and the increase in threats to research integrity. 24 00:01:05.969 --> 00:01:07.349 - Here's that conversation. 25 00:01:14.415 --> 00:01:14.915 - Hi, 26 00:01:16.254 --> 00:01:16.754 - Antonia. 27 00:01:17.055 --> 00:01:18.194 - Welcome to the podcast. 28 00:01:18.974 --> 00:01:21.075 - Hi, Hamish. Thank you for having me. 29 00:01:21.454 --> 00:01:22.274 - So, Antonia, 30 00:01:22.734 --> 00:01:26.034 - global research output is growing at an astonishing 31 00:01:26.254 --> 00:01:29.060 - rate. Is the current peer review system still 32 00:01:29.060 --> 00:01:31.379 - fit for purpose when it comes to dealing 33 00:01:31.379 --> 00:01:33.239 - with this volume of information? 34 00:01:33.939 --> 00:01:36.840 - Or do we need to develop alternative methods 35 00:01:37.060 --> 00:01:37.879 - for disseminating 36 00:01:38.579 --> 00:01:40.119 - high quality information? 37 00:01:41.844 --> 00:01:44.104 - That's a great question. I mean, peer review 38 00:01:44.165 --> 00:01:46.104 - for me is still absolutely fundamental 39 00:01:46.484 --> 00:01:48.025 - to how we assess research, 40 00:01:48.564 --> 00:01:50.484 - but we can't take it for granted. So 41 00:01:50.484 --> 00:01:53.364 - we need to keep strengthening it, making smart 42 00:01:53.364 --> 00:01:54.344 - use of technology 43 00:01:55.030 --> 00:01:57.270 - to ensure that it remains fit for the 44 00:01:57.270 --> 00:01:57.770 - future. 45 00:01:58.230 --> 00:02:00.730 - Because as you point out, research outputs continue 46 00:02:00.790 --> 00:02:02.650 - to grow at this amazing rate, 47 00:02:02.950 --> 00:02:06.409 - and that's driven by an increased research activity. 48 00:02:06.549 --> 00:02:09.050 - Yes. But it's also as a result of 49 00:02:09.349 --> 00:02:10.490 - AI tools making 50 00:02:11.165 --> 00:02:12.625 - generating research papers 51 00:02:13.004 --> 00:02:14.064 - faster and easier 52 00:02:14.444 --> 00:02:16.764 - because it can accelerate pretty much every stage 53 00:02:16.764 --> 00:02:18.224 - of the research workflow 54 00:02:18.764 --> 00:02:22.064 - from literature review through to drafting and revision. 55 00:02:22.844 --> 00:02:23.344 - So 56 00:02:23.760 --> 00:02:26.340 - I I think for for me, in an 57 00:02:26.480 --> 00:02:29.199 - AI era, if you like, peer review becomes 58 00:02:29.199 --> 00:02:31.219 - even more important, not less. 59 00:02:31.760 --> 00:02:34.740 - Because as information becomes easier to generate, 60 00:02:35.520 --> 00:02:37.794 - trust is harder to come by and trust 61 00:02:37.794 --> 00:02:40.055 - in that content is harder to establish. 62 00:02:40.835 --> 00:02:41.235 - So, 63 00:02:41.875 --> 00:02:43.474 - I think it's fair to say that peer 64 00:02:43.474 --> 00:02:45.634 - review is one of the strongest trust trust 65 00:02:45.634 --> 00:02:46.694 - signals available, 66 00:02:47.715 --> 00:02:48.210 - but, 67 00:02:48.930 --> 00:02:51.170 - it needs to evolve so that it becomes 68 00:02:51.170 --> 00:02:52.950 - more visible and machine readable 69 00:02:53.569 --> 00:02:56.370 - so that AI systems can recognize that it's 70 00:02:56.370 --> 00:02:56.870 - authoritative 71 00:02:57.250 --> 00:02:57.750 - research. 72 00:02:58.370 --> 00:03:00.370 - And that's gonna help protect against one of 73 00:03:00.370 --> 00:03:03.330 - the biggest risks of the AI era, which 74 00:03:03.330 --> 00:03:03.764 - is 75 00:03:04.164 --> 00:03:04.824 - the amplification 76 00:03:05.125 --> 00:03:06.745 - of errors and misinformation. 77 00:03:08.004 --> 00:03:09.525 - So you can see how easy it is 78 00:03:09.525 --> 00:03:12.264 - for AI to generate summaries and make connections 79 00:03:13.284 --> 00:03:14.504 - at rapid speeds, 80 00:03:14.884 --> 00:03:16.620 - but it can't confer trust. 81 00:03:17.659 --> 00:03:19.519 - So if peer review 82 00:03:19.979 --> 00:03:22.459 - can do something that AI can't, which is 83 00:03:22.459 --> 00:03:25.120 - that human expert judgment and accountability, 84 00:03:26.299 --> 00:03:28.479 - then that becomes ever more important. 85 00:03:29.019 --> 00:03:30.539 - Because at the end of the day, the 86 00:03:30.539 --> 00:03:31.759 - editorial standards 87 00:03:32.885 --> 00:03:35.605 - through peer review are applied by humans, by 88 00:03:35.605 --> 00:03:36.905 - real named individuals. 89 00:03:38.085 --> 00:03:39.844 - So I think, you know, in the past, 90 00:03:39.844 --> 00:03:42.325 - peer review helped determine whether a paper merited 91 00:03:42.325 --> 00:03:42.825 - publication, 92 00:03:43.764 --> 00:03:45.764 - but going forward, it's gonna be more of 93 00:03:45.764 --> 00:03:48.310 - a a a signal, a trust marker to 94 00:03:48.310 --> 00:03:49.770 - both humans and machines 95 00:03:50.469 --> 00:03:52.409 - that that research has been independently 96 00:03:52.710 --> 00:03:53.210 - scrutinized 97 00:03:53.990 --> 00:03:54.889 - and is therefore 98 00:03:55.430 --> 00:03:57.990 - of of the caliber that you would want 99 00:03:57.990 --> 00:03:59.270 - it to be if you were going to 100 00:03:59.270 --> 00:03:59.925 - use it. 101 00:04:00.324 --> 00:04:02.004 - So I don't think the challenge is as 102 00:04:02.004 --> 00:04:04.504 - much about whether peer review remains relevant. 103 00:04:04.805 --> 00:04:05.944 - It's more about 104 00:04:06.324 --> 00:04:07.865 - how we can evolve it 105 00:04:08.324 --> 00:04:11.525 - to develop from the, publishing process into more 106 00:04:11.525 --> 00:04:12.985 - of a kind of trusted infrastructure. 107 00:04:13.284 --> 00:04:14.905 - So it's not a one off event, 108 00:04:15.469 --> 00:04:16.850 - but it's something that 109 00:04:17.230 --> 00:04:20.509 - lives and creates this this trust around the 110 00:04:20.509 --> 00:04:21.569 - research itself. 111 00:04:22.350 --> 00:04:23.790 - I think it's safe to say that when 112 00:04:23.790 --> 00:04:25.550 - you when you speak to people in the 113 00:04:25.550 --> 00:04:29.625 - academic community, you know, research physicists, there is 114 00:04:29.625 --> 00:04:32.425 - huge support for peer review, isn't it? I 115 00:04:32.425 --> 00:04:34.925 - mean, I don't think I've never heard any 116 00:04:35.305 --> 00:04:35.805 - complaints 117 00:04:36.185 --> 00:04:38.425 - about, oh, we you know, about peer review 118 00:04:38.425 --> 00:04:40.904 - from the community. Maybe they want, you know, 119 00:04:40.904 --> 00:04:42.199 - it more streamlined 120 00:04:42.579 --> 00:04:43.079 - or, 121 00:04:43.939 --> 00:04:45.079 - you know, more open. 122 00:04:45.459 --> 00:04:47.540 - But I think it it does remain the 123 00:04:47.540 --> 00:04:49.620 - the sort of the backbone of science, doesn't 124 00:04:49.620 --> 00:04:51.379 - it? I don't I don't think anyone's come 125 00:04:51.379 --> 00:04:53.220 - up with anything better. I think that's the 126 00:04:53.220 --> 00:04:55.560 - point. I mean, it has its flaws, and 127 00:04:55.620 --> 00:04:58.375 - overload is a big problem for peer reviewers. 128 00:04:58.754 --> 00:05:01.074 - You know, as research grows, there's more that 129 00:05:01.074 --> 00:05:03.574 - needs to be validated and filtered. 130 00:05:03.954 --> 00:05:06.615 - So I think that problem is is endemic 131 00:05:06.675 --> 00:05:09.175 - and getting worse. So, again, 132 00:05:10.649 --> 00:05:13.449 - AI can help, not just AI, but there 133 00:05:13.449 --> 00:05:15.529 - are many tools available now that can take 134 00:05:15.529 --> 00:05:18.250 - some of the the burden away from peer 135 00:05:18.250 --> 00:05:20.009 - reviewers so that they can do the the 136 00:05:20.009 --> 00:05:22.009 - the kind of human judgment piece and look 137 00:05:22.009 --> 00:05:22.750 - at the originality 138 00:05:23.605 --> 00:05:25.605 - and the kind of credibility of the science 139 00:05:25.605 --> 00:05:28.245 - rather than having to to sift through some 140 00:05:28.245 --> 00:05:30.725 - of the technical aspects or look for for 141 00:05:30.725 --> 00:05:33.524 - research integrity problems, which, you know, I know 142 00:05:33.524 --> 00:05:35.205 - we're we're gonna go on to talk about 143 00:05:35.205 --> 00:05:37.110 - a bit more. So I think I think 144 00:05:37.110 --> 00:05:38.410 - peer review remains 145 00:05:38.790 --> 00:05:39.290 - foundational, 146 00:05:40.389 --> 00:05:42.550 - but it has to evolve, and it has 147 00:05:42.550 --> 00:05:43.050 - to 148 00:05:43.430 --> 00:05:46.150 - has to remain relevant in a world where 149 00:05:46.150 --> 00:05:47.449 - so much more of, 150 00:05:48.150 --> 00:05:48.650 - research 151 00:05:49.425 --> 00:05:52.464 - creation, but also research discovery is gonna happen 152 00:05:52.464 --> 00:05:53.204 - by machine. 153 00:05:54.224 --> 00:05:56.464 - So that would be my my view at 154 00:05:56.464 --> 00:05:59.664 - the moment. So so peer reviewing something produced 155 00:05:59.664 --> 00:06:00.485 - by a machine, 156 00:06:01.040 --> 00:06:02.480 - is that, I mean, is that something that 157 00:06:02.480 --> 00:06:04.079 - we're gonna have to start thinking about? I 158 00:06:04.079 --> 00:06:05.379 - suppose it is. Aspects 159 00:06:05.920 --> 00:06:08.520 - of the peer review could be produced by 160 00:06:08.520 --> 00:06:09.139 - a machine, 161 00:06:09.680 --> 00:06:11.839 - but led by the human. I mean, the 162 00:06:11.839 --> 00:06:14.800 - human will be operating the machine, interpreting the 163 00:06:14.800 --> 00:06:15.860 - machine's results. 164 00:06:16.504 --> 00:06:17.324 - But I think, 165 00:06:17.865 --> 00:06:20.665 - the point that that where peer review needs 166 00:06:20.665 --> 00:06:23.004 - to evolve is that at the moment, 167 00:06:23.545 --> 00:06:25.625 - you see an article in a journal and 168 00:06:25.625 --> 00:06:28.425 - you know by virtue of the brand or 169 00:06:28.425 --> 00:06:31.064 - the publisher behind it that that content has 170 00:06:31.064 --> 00:06:33.300 - been peer reviewed, but there isn't actually anything. 171 00:06:33.520 --> 00:06:35.620 - There's no sort of kite mark that says 172 00:06:36.000 --> 00:06:37.540 - that is a peer reviewed article. 173 00:06:37.840 --> 00:06:39.939 - So when the content from that article 174 00:06:40.639 --> 00:06:43.279 - becomes unlocked from the article as it might 175 00:06:43.279 --> 00:06:45.454 - do if you if you create an AI 176 00:06:45.454 --> 00:06:47.634 - summary in a in a question you ask 177 00:06:48.095 --> 00:06:49.235 - Claude or ChatGPT, 178 00:06:50.095 --> 00:06:52.814 - then there is no peer review trust marker 179 00:06:52.814 --> 00:06:54.274 - that travels with that 180 00:06:54.814 --> 00:06:55.314 - content, 181 00:06:55.855 --> 00:06:58.259 - let alone the whole article. So we need 182 00:06:58.259 --> 00:07:00.120 - to work out how we codify 183 00:07:00.899 --> 00:07:04.019 - that peer review trust marker into the content 184 00:07:04.019 --> 00:07:05.160 - so that it travels 185 00:07:06.339 --> 00:07:09.240 - with the content through machine to machine interfaces. 186 00:07:10.419 --> 00:07:12.419 - And that's really important, isn't it, 187 00:07:13.074 --> 00:07:15.154 - These days, I suppose, because there's just so 188 00:07:15.154 --> 00:07:17.095 - much information being produced 189 00:07:17.634 --> 00:07:20.694 - and and modified and transcribed and translated 190 00:07:21.314 --> 00:07:24.355 - and moved around. Absolutely. And and that, I 191 00:07:24.355 --> 00:07:27.419 - suppose, has opened the door to fraud. I 192 00:07:27.419 --> 00:07:30.240 - mean, I think fraud in, scholarly publishing 193 00:07:30.779 --> 00:07:32.620 - is is something that you're you're seeing more 194 00:07:32.620 --> 00:07:33.519 - and more of 195 00:07:33.819 --> 00:07:34.879 - year on year. 196 00:07:36.060 --> 00:07:37.279 - So so how is 197 00:07:37.660 --> 00:07:38.639 - IOP publishing, 198 00:07:39.660 --> 00:07:43.675 - dealing with the submission of fraud fraudulent papers? 199 00:07:43.675 --> 00:07:44.735 - How how do you, 200 00:07:45.354 --> 00:07:47.375 - identify them and and reject them? 201 00:07:48.555 --> 00:07:49.914 - Yeah. I mean, I I agree. I think 202 00:07:49.914 --> 00:07:51.995 - the rise in fraudulent papers is one one 203 00:07:51.995 --> 00:07:53.134 - of the most concerning 204 00:07:53.995 --> 00:07:55.774 - issues for the scholarly 205 00:07:56.430 --> 00:08:00.189 - publishing ecosystem. Well, the scholarly ecosystem, frankly, because 206 00:08:00.189 --> 00:08:01.490 - it threatens not only 207 00:08:01.870 --> 00:08:04.850 - the individual journals, but the publishers behind them, 208 00:08:05.790 --> 00:08:08.110 - and also the confidence in the research record 209 00:08:08.110 --> 00:08:09.170 - itself and therefore 210 00:08:09.870 --> 00:08:12.915 - how much trust society puts into science. 211 00:08:13.295 --> 00:08:16.115 - So it's it's absolutely critical that we 212 00:08:16.574 --> 00:08:18.275 - address this, and IOPPAEC's 213 00:08:19.215 --> 00:08:22.014 - always has done, but more so now than 214 00:08:22.014 --> 00:08:24.335 - ever has put research integrity at the front 215 00:08:24.335 --> 00:08:25.715 - of its strategic priorities. 216 00:08:26.379 --> 00:08:28.800 - It fits with our purpose led publishing 217 00:08:29.180 --> 00:08:29.680 - mission. 218 00:08:30.939 --> 00:08:33.679 - So our response is a combination of 219 00:08:34.059 --> 00:08:34.960 - human expertise 220 00:08:35.340 --> 00:08:35.840 - and 221 00:08:36.299 --> 00:08:38.220 - tooling, as we've just been talking about. So 222 00:08:38.220 --> 00:08:40.915 - we've invested since I became CEO. The number 223 00:08:40.915 --> 00:08:42.595 - of people that work in our peer review 224 00:08:42.595 --> 00:08:45.315 - teams and in our research integrity teams has 225 00:08:45.315 --> 00:08:46.774 - grown very significantly. 226 00:08:47.475 --> 00:08:49.415 - So there are there are people, 227 00:08:50.115 --> 00:08:52.754 - experts at the heart of trying to make 228 00:08:52.754 --> 00:08:54.835 - sure that we maintain the trust in that 229 00:08:54.835 --> 00:08:56.370 - scientific record through 230 00:08:56.850 --> 00:08:59.029 - catching the fraudulent content early. 231 00:08:59.649 --> 00:09:01.409 - But the tooling is, 232 00:09:02.529 --> 00:09:04.610 - is helping us spot patterns that the human 233 00:09:04.610 --> 00:09:06.230 - eye probably wouldn't spot. 234 00:09:06.769 --> 00:09:09.409 - So you probably heard about tortured phrases. You've 235 00:09:09.409 --> 00:09:12.149 - heard obviously, plagiarism has always been around. But 236 00:09:12.495 --> 00:09:14.995 - with the with the ability of machines 237 00:09:15.375 --> 00:09:16.674 - now, we can, 238 00:09:18.815 --> 00:09:20.975 - weed out we can spot and weed out 239 00:09:20.975 --> 00:09:23.934 - this content much more easily than we used 240 00:09:23.934 --> 00:09:24.269 - to, 241 00:09:24.990 --> 00:09:26.910 - used to be able to. So I would 242 00:09:26.910 --> 00:09:29.149 - say that AI, again, is a positive in 243 00:09:29.149 --> 00:09:30.350 - that regard because it's, 244 00:09:31.389 --> 00:09:33.870 - helping us get better at this. In many 245 00:09:33.870 --> 00:09:36.670 - cases, we're using AI to fight AI because 246 00:09:36.670 --> 00:09:38.670 - many of much of the fraudulent content has 247 00:09:38.670 --> 00:09:41.285 - been generated using AI because it makes it 248 00:09:41.285 --> 00:09:43.605 - so much easier to just, you know, push 249 00:09:43.605 --> 00:09:44.825 - out a paper that 250 00:09:45.605 --> 00:09:47.065 - that would have taken years 251 00:09:47.365 --> 00:09:48.904 - to prepare in the past, 252 00:09:49.605 --> 00:09:50.745 - and fraud, 253 00:09:51.045 --> 00:09:54.264 - create fraudulent citations and fraudulent data, etcetera. 254 00:09:55.009 --> 00:09:57.570 - So the problem is growing, but we have 255 00:09:57.809 --> 00:09:59.250 - on the on the other side of it, 256 00:09:59.250 --> 00:10:02.129 - we have better armoury, if you like, to 257 00:10:02.129 --> 00:10:02.870 - fight it. 258 00:10:03.730 --> 00:10:05.750 - So it's a combination of of both. 259 00:10:08.024 --> 00:10:09.865 - I think the important thing, as we just 260 00:10:09.865 --> 00:10:12.665 - said before, is that the humans are still 261 00:10:12.665 --> 00:10:13.805 - in charge of 262 00:10:14.184 --> 00:10:16.425 - the peer review end to end process well, 263 00:10:16.425 --> 00:10:18.504 - the whole publishing process, but the peer review 264 00:10:18.504 --> 00:10:19.004 - process 265 00:10:19.625 --> 00:10:21.465 - because that then means that, 266 00:10:21.865 --> 00:10:24.550 - you know, the flags that these tools 267 00:10:24.930 --> 00:10:27.330 - are sending up when they look through the 268 00:10:27.330 --> 00:10:27.830 - content 269 00:10:28.370 --> 00:10:30.529 - are being checked because, you know, these people's 270 00:10:30.610 --> 00:10:32.610 - this is people's reputation, and if we get 271 00:10:32.610 --> 00:10:33.269 - it wrong, 272 00:10:33.810 --> 00:10:36.304 - that's, you know, that's on us. So making 273 00:10:36.304 --> 00:10:38.065 - sure that we have human judgment in the 274 00:10:38.065 --> 00:10:39.745 - loop and that the time that the humans 275 00:10:39.745 --> 00:10:42.245 - spend is on that critical thinking and judgment 276 00:10:42.304 --> 00:10:44.065 - rather than on trying to spot the things 277 00:10:44.065 --> 00:10:45.044 - in the first place. 278 00:10:46.225 --> 00:10:47.684 - And we've also engaged, 279 00:10:48.304 --> 00:10:50.705 - at IOP, along with many other publishers with, 280 00:10:51.559 --> 00:10:53.179 - various different collective actions. 281 00:10:53.959 --> 00:10:57.100 - So there's, something called the STM Integrity Hub. 282 00:10:57.399 --> 00:11:00.039 - So it's combining signals not just from, you 283 00:11:00.039 --> 00:11:02.600 - know, our corpus of content, but from Springer 284 00:11:02.600 --> 00:11:04.620 - Nature's or from Wiley's or from, 285 00:11:05.665 --> 00:11:06.865 - American Institute of Publishing, 286 00:11:07.425 --> 00:11:09.665 - American Institute of Physics Publishing, you know, you 287 00:11:09.665 --> 00:11:10.404 - name it. 288 00:11:10.785 --> 00:11:12.325 - All of publishers are being, 289 00:11:13.825 --> 00:11:16.965 - attacked, if you like, by these bad actors 290 00:11:17.024 --> 00:11:19.925 - trying to infiltrate the the system with fraudulent 291 00:11:19.985 --> 00:11:22.100 - papers. So if we can read the signals 292 00:11:22.100 --> 00:11:23.860 - across all of the content, then if we 293 00:11:23.860 --> 00:11:24.759 - reject something 294 00:11:25.139 --> 00:11:26.360 - at IAPP, hopefully, 295 00:11:26.740 --> 00:11:28.179 - there aren't ways in which it can be 296 00:11:28.179 --> 00:11:29.799 - then picked up by another publisher. 297 00:11:30.259 --> 00:11:32.019 - So we keep it out once, and we 298 00:11:32.019 --> 00:11:33.779 - keep it out for for all, if you 299 00:11:33.779 --> 00:11:34.034 - like. 300 00:11:35.074 --> 00:11:36.674 - And the other thing I think that's really 301 00:11:36.674 --> 00:11:37.174 - important, 302 00:11:37.714 --> 00:11:39.554 - and this is a difficult one because you 303 00:11:39.554 --> 00:11:42.774 - have organizations like Clarivate who index journals 304 00:11:43.394 --> 00:11:44.115 - and and, 305 00:11:44.754 --> 00:11:47.634 - basically manage the citation reporting of journals to 306 00:11:47.634 --> 00:11:50.139 - give the the impact factors that are so 307 00:11:50.459 --> 00:11:50.959 - unfortunately, 308 00:11:51.259 --> 00:11:52.879 - remain so important in 309 00:11:53.500 --> 00:11:54.080 - in research, 310 00:11:55.419 --> 00:11:56.480 - in research recognition. 311 00:11:57.740 --> 00:12:00.159 - But Clarivate and others are 312 00:12:00.475 --> 00:12:02.715 - trying to clean up the scholarly record too. 313 00:12:02.715 --> 00:12:04.654 - And so if you're found to be publishing 314 00:12:04.715 --> 00:12:07.195 - too much fraudulent content, then your journal can 315 00:12:07.195 --> 00:12:08.014 - be delisted. 316 00:12:09.034 --> 00:12:10.975 - And sometimes that can be a disincentive 317 00:12:11.355 --> 00:12:12.495 - for trying to 318 00:12:13.110 --> 00:12:15.450 - clear up the record because if you're penalized 319 00:12:15.669 --> 00:12:18.149 - for doing this, then maybe you don't want 320 00:12:18.149 --> 00:12:20.070 - to go looking to see if you've published 321 00:12:20.070 --> 00:12:22.230 - stuff that, you know, maybe you shouldn't have 322 00:12:22.230 --> 00:12:23.669 - done in the past. And it's fair to 323 00:12:23.669 --> 00:12:25.190 - say that, you know, maybe we didn't see 324 00:12:25.190 --> 00:12:27.445 - these problems quite as much because we didn't 325 00:12:27.445 --> 00:12:28.884 - have the tooling in the past to be 326 00:12:28.884 --> 00:12:30.504 - able to ask to do that. So, 327 00:12:31.205 --> 00:12:33.764 - I think, you know, transparency and and correcting 328 00:12:33.764 --> 00:12:36.264 - the record's really important. But at the moment, 329 00:12:36.325 --> 00:12:37.924 - we've got quite a lot of sticks and 330 00:12:37.924 --> 00:12:40.024 - perhaps not enough carrots for publishers 331 00:12:40.659 --> 00:12:41.480 - to do that. 332 00:12:41.940 --> 00:12:43.700 - But one of the things I'm most proud 333 00:12:43.700 --> 00:12:45.940 - of that we've done at IAPP, and that's 334 00:12:45.940 --> 00:12:47.940 - back to our values and our and our 335 00:12:47.940 --> 00:12:48.440 - purpose, 336 00:12:49.460 --> 00:12:52.179 - is that when we retract an open access 337 00:12:52.179 --> 00:12:54.820 - article where someone's paid an article publishing charge 338 00:12:54.820 --> 00:12:55.639 - or an APC, 339 00:12:56.754 --> 00:12:58.855 - we actually donate that retracted 340 00:12:59.554 --> 00:13:03.235 - APC payment to an organization called Research for 341 00:13:03.235 --> 00:13:03.735 - Life 342 00:13:04.195 --> 00:13:06.754 - that helps, people in lower and middle income 343 00:13:06.754 --> 00:13:09.154 - countries to gain access and and and better 344 00:13:09.154 --> 00:13:09.654 - publishing, 345 00:13:10.834 --> 00:13:11.334 - opportunities. 346 00:13:13.290 --> 00:13:15.049 - So I think that's the right thing to 347 00:13:15.049 --> 00:13:17.629 - do, but it really, I think, demonstrates that 348 00:13:17.690 --> 00:13:20.970 - for IOPP, putting research integrity above revenue and 349 00:13:20.970 --> 00:13:23.450 - profit is is what we is what we 350 00:13:23.450 --> 00:13:25.710 - need to do. And do you know, 351 00:13:26.514 --> 00:13:28.754 - or do publishers know, you know, what the 352 00:13:28.754 --> 00:13:29.574 - main motivation 353 00:13:30.355 --> 00:13:32.115 - behind fraud is? I mean, I would have 354 00:13:32.115 --> 00:13:34.034 - assumed it's sort of publish and pair or 355 00:13:34.034 --> 00:13:36.754 - perish, and someone's desperate to, you know, by 356 00:13:36.754 --> 00:13:38.355 - the end of the year, they need a 357 00:13:38.355 --> 00:13:40.834 - a publication, and so they they just make 358 00:13:40.834 --> 00:13:41.279 - it up. 359 00:13:41.840 --> 00:13:44.399 - I mean, I'm guessing that that's probably most 360 00:13:44.399 --> 00:13:47.779 - of the fraud. But has anybody identified anything, 361 00:13:47.920 --> 00:13:49.139 - you know, more nefarious? 362 00:13:49.759 --> 00:13:51.279 - You know, because we live in a in 363 00:13:51.279 --> 00:13:53.220 - a world now where we have to be 364 00:13:54.205 --> 00:13:56.764 - worried about misinformation coming from a number of 365 00:13:56.764 --> 00:13:58.465 - different sources. I mean, 366 00:13:59.085 --> 00:14:00.304 - has anybody identified 367 00:14:00.605 --> 00:14:01.105 - any, 368 00:14:01.644 --> 00:14:03.904 - you know, sort of organized misinformation? 369 00:14:04.845 --> 00:14:07.245 - You know, either to discredit a publisher, as 370 00:14:07.245 --> 00:14:07.904 - you said, 371 00:14:08.330 --> 00:14:11.230 - or, I don't know, discredit a field or 372 00:14:11.690 --> 00:14:14.889 - discredit a a concept in a in science, 373 00:14:14.889 --> 00:14:17.850 - or or or maybe it's mostly just people 374 00:14:17.850 --> 00:14:18.350 - desperate, 375 00:14:19.210 --> 00:14:21.309 - to get something on their publication list. 376 00:14:21.954 --> 00:14:23.875 - Yeah. I I think I think you're absolutely 377 00:14:23.875 --> 00:14:26.595 - right. It's the publish or perish mentality, and, 378 00:14:26.595 --> 00:14:28.214 - you know, that's why you're here 379 00:14:28.595 --> 00:14:29.654 - repeated calls 380 00:14:30.195 --> 00:14:31.894 - to look again at the way 381 00:14:32.834 --> 00:14:33.815 - research is, 382 00:14:34.434 --> 00:14:35.975 - measured and how researchers 383 00:14:36.355 --> 00:14:36.855 - are, 384 00:14:37.789 --> 00:14:40.669 - rewarded and how their careers progress because it 385 00:14:40.669 --> 00:14:42.129 - does all center around 386 00:14:43.149 --> 00:14:45.230 - the quantity as well as as well as 387 00:14:45.230 --> 00:14:48.370 - the quality, but increasingly quantity of, 388 00:14:48.830 --> 00:14:50.909 - publications that a researcher has. So it does 389 00:14:50.909 --> 00:14:53.464 - put undue pressure on them. So I think 390 00:14:53.464 --> 00:14:56.204 - the vast majority is being driven by 391 00:14:56.985 --> 00:14:57.644 - by those 392 00:14:58.264 --> 00:15:00.845 - those sort of base needs for a researcher 393 00:15:00.904 --> 00:15:02.125 - to have to to publish, 394 00:15:02.504 --> 00:15:04.664 - to progress in their career. But it has 395 00:15:04.664 --> 00:15:07.690 - become much more organised and it's become sort 396 00:15:07.690 --> 00:15:09.789 - of industrialized, if you like, because, 397 00:15:11.370 --> 00:15:13.049 - you know, it might be one or two 398 00:15:13.049 --> 00:15:15.610 - bad actors in the past, but it's become 399 00:15:15.610 --> 00:15:18.089 - much more visible. And there are companies out 400 00:15:18.089 --> 00:15:19.549 - there now selling, 401 00:15:20.125 --> 00:15:21.725 - you know, the chance to add your name 402 00:15:21.725 --> 00:15:22.785 - onto a paper 403 00:15:23.165 --> 00:15:25.245 - or, you know, as a a a number 404 00:15:25.245 --> 00:15:27.644 - of citations to a to a range of 405 00:15:27.644 --> 00:15:28.945 - different papers, etcetera. 406 00:15:29.404 --> 00:15:31.884 - And that's because people can see they can 407 00:15:31.884 --> 00:15:33.690 - make money out of it. And that's not 408 00:15:33.690 --> 00:15:34.190 - researchers 409 00:15:35.210 --> 00:15:36.190 - doing that. That's, 410 00:15:36.730 --> 00:15:40.009 - you know, tricksters seeing an opportunity to to 411 00:15:40.009 --> 00:15:42.009 - make quite big money, and they're advertising in 412 00:15:42.009 --> 00:15:44.009 - plain sight on a lot of the social 413 00:15:44.009 --> 00:15:44.910 - media platforms. 414 00:15:45.674 --> 00:15:47.195 - And so, you know, even if you're a 415 00:15:47.195 --> 00:15:49.375 - researcher that hadn't even thought about doing that, 416 00:15:49.674 --> 00:15:51.835 - you know, now it's being promoted to you 417 00:15:51.835 --> 00:15:53.995 - as something you could do. I think it's 418 00:15:53.995 --> 00:15:56.075 - probably tempting more people than would have done 419 00:15:56.075 --> 00:15:58.075 - it in the past because it's being done 420 00:15:58.075 --> 00:15:59.889 - at scale. So it's sort of a a 421 00:15:59.889 --> 00:16:02.470 - form of organized crime, really. Absolutely. 422 00:16:03.410 --> 00:16:04.950 - Oh, that that is unfortunate. 423 00:16:05.410 --> 00:16:07.250 - Yeah. But I suppose that's the way it 424 00:16:07.250 --> 00:16:10.210 - goes when when money and and prestige and 425 00:16:10.690 --> 00:16:11.429 - is involved. 426 00:16:12.565 --> 00:16:14.585 - You you've mentioned artificial intelligence, 427 00:16:15.445 --> 00:16:16.585 - a few times already, 428 00:16:17.205 --> 00:16:19.125 - and, I think it's safe to say that 429 00:16:19.125 --> 00:16:21.304 - it offers great challenges and tremendous 430 00:16:22.004 --> 00:16:22.504 - opportunities 431 00:16:22.965 --> 00:16:24.264 - for those who create 432 00:16:24.779 --> 00:16:27.039 - and document scientific knowledge. 433 00:16:27.419 --> 00:16:30.320 - How how is AI changing scholarly publishing? 434 00:16:31.659 --> 00:16:33.419 - Yeah. You're right. I mean, I've started the 435 00:16:33.419 --> 00:16:35.980 - podcast off talking about AI from the get 436 00:16:35.980 --> 00:16:37.215 - go, and that's because it's, 437 00:16:38.175 --> 00:16:39.555 - a pervasive force 438 00:16:40.014 --> 00:16:42.254 - that touches every aspect of our business. And 439 00:16:42.254 --> 00:16:44.274 - I would argue pretty much every business. 440 00:16:45.934 --> 00:16:47.774 - And, you know, I sort of lost track 441 00:16:47.774 --> 00:16:49.774 - of how many times colleagues and actually board 442 00:16:49.774 --> 00:16:52.440 - members as well have have asked me, what's 443 00:16:52.440 --> 00:16:52.940 - IAPP's 444 00:16:53.399 --> 00:16:54.299 - AI strategy? 445 00:16:55.240 --> 00:16:57.080 - And I I actually push back on that. 446 00:16:57.080 --> 00:16:59.159 - I don't think that's the right question to 447 00:16:59.159 --> 00:17:01.799 - be asking. The question is is really how 448 00:17:01.799 --> 00:17:04.619 - does AI shape everything that we do? 449 00:17:05.065 --> 00:17:06.744 - You know, we couldn't really have an AI 450 00:17:06.744 --> 00:17:08.825 - strategy because it's so many different things. It's 451 00:17:08.825 --> 00:17:11.545 - it's a technology. You'd never have an Excel 452 00:17:11.545 --> 00:17:13.545 - strategy or you'd never have a you know, 453 00:17:13.545 --> 00:17:14.825 - in the past, to be fair, we did 454 00:17:14.825 --> 00:17:16.984 - used to have digital strategies, and you had 455 00:17:16.984 --> 00:17:18.984 - people with titles that were digital this or 456 00:17:18.984 --> 00:17:21.460 - digital that. Now you've got people in companies 457 00:17:21.759 --> 00:17:24.160 - with titles AI this or AI that, but, 458 00:17:24.160 --> 00:17:24.660 - ultimately, 459 00:17:25.440 --> 00:17:27.859 - it's an enabler of lots of different things. 460 00:17:28.320 --> 00:17:30.960 - So we're in the process of developing our 461 00:17:30.960 --> 00:17:32.559 - new 2027 462 00:17:32.559 --> 00:17:33.700 - to 2030 463 00:17:34.184 --> 00:17:35.085 - company strategy. 464 00:17:36.024 --> 00:17:38.044 - And so we've approached that by considering 465 00:17:38.505 --> 00:17:41.325 - the implications of AI across all dimensions 466 00:17:41.704 --> 00:17:42.365 - of our 467 00:17:43.065 --> 00:17:44.764 - of our thinking and our activities. 468 00:17:45.625 --> 00:17:47.085 - So it's a source of disruption 469 00:17:47.659 --> 00:17:50.059 - to traditional publishing models. We've talked a little 470 00:17:50.059 --> 00:17:52.220 - bit about that. It's a driver of new 471 00:17:52.220 --> 00:17:52.720 - opportunities 472 00:17:53.099 --> 00:17:54.319 - for growth and, 473 00:17:54.940 --> 00:17:55.440 - differentiation, 474 00:17:56.539 --> 00:17:59.519 - and also a catalyst for step change improvements 475 00:17:59.659 --> 00:18:02.065 - in in how we operate. So for me, 476 00:18:02.065 --> 00:18:04.545 - the opportunities are very exciting and and pretty 477 00:18:04.545 --> 00:18:05.045 - significant. 478 00:18:06.384 --> 00:18:08.725 - And we talked a little bit about how 479 00:18:09.345 --> 00:18:10.244 - more discovery, 480 00:18:10.865 --> 00:18:14.305 - particularly research discovery, is being mediated by AI 481 00:18:14.305 --> 00:18:14.805 - systems. 482 00:18:16.029 --> 00:18:18.430 - So scholarly publishers really need to invest in 483 00:18:18.430 --> 00:18:20.769 - making our content much more machine readable, 484 00:18:21.470 --> 00:18:22.930 - discoverable, but also 485 00:18:23.390 --> 00:18:23.890 - attributable 486 00:18:25.150 --> 00:18:26.529 - in those kind of environments. 487 00:18:27.710 --> 00:18:29.009 - And so I think 488 00:18:29.365 --> 00:18:30.184 - for us 489 00:18:30.484 --> 00:18:33.464 - as as a publisher of domain specific, 490 00:18:33.924 --> 00:18:35.784 - high quality, trusted content, 491 00:18:36.724 --> 00:18:39.684 - the opportunity to license that content to make 492 00:18:39.684 --> 00:18:42.724 - sure it's discovered and integrated and monetized in 493 00:18:42.724 --> 00:18:45.149 - different ways, as I say, is quite exciting. 494 00:18:46.909 --> 00:18:49.230 - So I think if you think about the 495 00:18:49.230 --> 00:18:51.089 - traditional journal container 496 00:18:51.390 --> 00:18:52.929 - or even the article container, 497 00:18:54.029 --> 00:18:56.109 - AI creates an opportunity to break out of 498 00:18:56.109 --> 00:18:58.669 - that and enable our content to be found 499 00:18:58.669 --> 00:18:59.169 - in 500 00:18:59.684 --> 00:19:00.505 - totally different 501 00:19:01.125 --> 00:19:02.664 - workflows and use cases. 502 00:19:03.444 --> 00:19:05.924 - You know, we've we've never really been that 503 00:19:05.924 --> 00:19:07.144 - successful in reaching 504 00:19:07.684 --> 00:19:09.224 - corporates and professionals. 505 00:19:10.085 --> 00:19:12.325 - Content and the way it's produced and served 506 00:19:12.325 --> 00:19:14.325 - up at the moment is very academic in 507 00:19:14.325 --> 00:19:14.825 - nature. 508 00:19:15.820 --> 00:19:17.900 - But I think we can reach entirely new 509 00:19:17.900 --> 00:19:18.400 - audiences, 510 00:19:19.500 --> 00:19:20.000 - and 511 00:19:20.779 --> 00:19:22.640 - basically meet them in their workflows 512 00:19:23.420 --> 00:19:26.140 - through the benefits of of AI and machine 513 00:19:26.140 --> 00:19:26.880 - to machine 514 00:19:27.259 --> 00:19:27.759 - discovery. 515 00:19:28.525 --> 00:19:31.265 - So I think a new generation of, 516 00:19:32.525 --> 00:19:33.025 - of, 517 00:19:34.045 --> 00:19:34.944 - content delivery, 518 00:19:35.484 --> 00:19:36.704 - and I think the opportunity 519 00:19:37.005 --> 00:19:37.505 - to 520 00:19:38.525 --> 00:19:40.865 - improve the way that we deliver the services 521 00:19:40.950 --> 00:19:43.269 - we do through, you know, the productivity that 522 00:19:43.269 --> 00:19:44.710 - AI gives us in our day to day 523 00:19:44.710 --> 00:19:46.950 - work. I mean, that's not distinct to us 524 00:19:46.950 --> 00:19:48.970 - at IAPP or or to any 525 00:19:49.349 --> 00:19:51.690 - anyone in the publishing sector, but the productivity 526 00:19:51.909 --> 00:19:54.329 - that you can get through using AI tools, 527 00:19:54.775 --> 00:19:56.795 - I think, is is is fantastic. 528 00:19:57.894 --> 00:20:00.375 - So going forward, I think everyone that works 529 00:20:00.375 --> 00:20:00.875 - in 530 00:20:01.414 --> 00:20:03.755 - in our sector, everyone that works in IAPP 531 00:20:04.615 --> 00:20:06.394 - is going to need to be AI fluent. 532 00:20:06.855 --> 00:20:08.454 - We're going to need to make sure that 533 00:20:08.454 --> 00:20:09.115 - we're helping, 534 00:20:09.414 --> 00:20:10.315 - you know, colleagues 535 00:20:11.150 --> 00:20:12.450 - basically feel equipped, 536 00:20:13.230 --> 00:20:14.990 - with the skills, with the tools, but I 537 00:20:14.990 --> 00:20:16.690 - think most importantly, the mindset 538 00:20:17.069 --> 00:20:17.730 - as well 539 00:20:18.190 --> 00:20:20.529 - to you use AI effectively, 540 00:20:21.390 --> 00:20:21.890 - confidently, 541 00:20:22.429 --> 00:20:24.690 - and, of course, responsibly as well. 542 00:20:25.224 --> 00:20:25.724 - And 543 00:20:26.105 --> 00:20:27.724 - as a academic publisher, 544 00:20:28.345 --> 00:20:30.825 - I suppose we, you know, we're we're the 545 00:20:30.825 --> 00:20:33.625 - stewards of a huge amount of very high 546 00:20:33.625 --> 00:20:34.845 - quality information 547 00:20:35.304 --> 00:20:37.244 - that, I mean, in a way, the scientific 548 00:20:37.384 --> 00:20:38.605 - community has entrusted 549 00:20:39.464 --> 00:20:40.204 - us with. 550 00:20:40.509 --> 00:20:42.750 - I mean, do you feel, I mean, is 551 00:20:42.750 --> 00:20:43.970 - there a certain responsibility 552 00:20:44.269 --> 00:20:46.289 - that goes along with that to make sure 553 00:20:46.509 --> 00:20:48.529 - that that information is used 554 00:20:48.990 --> 00:20:49.490 - correctly 555 00:20:49.869 --> 00:20:50.609 - and appropriately 556 00:20:51.069 --> 00:20:53.890 - by AI systems and that it's not used, 557 00:20:54.775 --> 00:20:56.615 - well, for lack of a better word, in 558 00:20:56.615 --> 00:20:57.515 - a bad way 559 00:20:57.894 --> 00:20:59.974 - or in a way that's misleading. I mean, 560 00:20:59.974 --> 00:21:02.315 - is that is that the view that publishers 561 00:21:02.775 --> 00:21:03.515 - are taking? 562 00:21:04.535 --> 00:21:07.019 - It's absolutely the view we're taking. You know, 563 00:21:07.259 --> 00:21:09.500 - if we're going to license our content to 564 00:21:09.500 --> 00:21:11.039 - to AI tools, 565 00:21:11.820 --> 00:21:13.599 - AI agents, AI companies, 566 00:21:13.900 --> 00:21:15.339 - then we need to make sure that there 567 00:21:15.339 --> 00:21:17.500 - are the right guardrails around it and making 568 00:21:17.500 --> 00:21:18.400 - sure that the 569 00:21:18.779 --> 00:21:20.160 - attribution, the provenance, 570 00:21:21.524 --> 00:21:23.524 - the the ability to be able to update 571 00:21:23.524 --> 00:21:25.764 - the content if a retraction happens to make 572 00:21:25.764 --> 00:21:28.404 - sure that we're not repeating the same, you 573 00:21:28.404 --> 00:21:30.565 - know, bad information and that's feeding through all 574 00:21:30.565 --> 00:21:33.044 - of the systems. That's all easier said than 575 00:21:33.044 --> 00:21:35.940 - done, but that is the goal. And and 576 00:21:35.940 --> 00:21:37.240 - and absolutely, I think 577 00:21:37.619 --> 00:21:40.420 - what we've noticed more recently is that the 578 00:21:40.420 --> 00:21:41.240 - AI companies 579 00:21:41.700 --> 00:21:42.200 - themselves 580 00:21:42.740 --> 00:21:45.619 - are are are really interested in understanding how 581 00:21:45.619 --> 00:21:46.759 - to do this better. 582 00:21:47.275 --> 00:21:48.954 - So they've gone through their kind of brute 583 00:21:48.954 --> 00:21:52.075 - force stage of large language model training where 584 00:21:52.075 --> 00:21:54.394 - they just wanted to train the machines on 585 00:21:54.394 --> 00:21:56.575 - everything that they could get their hands on. 586 00:21:56.795 --> 00:21:57.855 - And frankly, they, 587 00:21:58.394 --> 00:22:00.480 - you know, the the I think there's evidence 588 00:22:00.480 --> 00:22:02.240 - that they've scraped a lot of our content 589 00:22:02.240 --> 00:22:04.019 - without permission to do so, 590 00:22:04.559 --> 00:22:06.640 - but many of them are now realizing that, 591 00:22:06.640 --> 00:22:09.599 - you know, their users aren't necessarily trusting the 592 00:22:09.599 --> 00:22:12.500 - answers that are being given by their tools. 593 00:22:13.015 --> 00:22:14.694 - So they need to rectify that or they're 594 00:22:14.694 --> 00:22:17.414 - not gonna be commercially successful. People won't want 595 00:22:17.414 --> 00:22:19.095 - to continue to use the tools. So many 596 00:22:19.095 --> 00:22:20.615 - of them going back to the drawing board 597 00:22:20.615 --> 00:22:22.634 - and thinking about, oh, okay. 598 00:22:23.015 --> 00:22:25.015 - You know, Reddit content is not the same 599 00:22:25.015 --> 00:22:25.674 - as IOPP 600 00:22:26.569 --> 00:22:28.589 - journal content. We need to 601 00:22:29.369 --> 00:22:30.829 - treat this content differently. 602 00:22:31.609 --> 00:22:34.269 - It has higher value in terms of that 603 00:22:34.410 --> 00:22:37.069 - trust marking that I was talking about earlier. 604 00:22:37.609 --> 00:22:40.109 - So I I think there's a coming together 605 00:22:40.329 --> 00:22:42.994 - and a an evolution of what started off 606 00:22:42.994 --> 00:22:45.575 - very much as a as a pure licensing 607 00:22:46.514 --> 00:22:48.375 - sort of market to 608 00:22:48.755 --> 00:22:49.255 - partnering. 609 00:22:49.714 --> 00:22:51.554 - So, you know, the job we do is 610 00:22:51.554 --> 00:22:53.815 - validation and filtration as publishers. 611 00:22:54.789 --> 00:22:56.789 - That's a job that, at the moment, looks 612 00:22:56.789 --> 00:22:58.730 - like traditional peer review. 613 00:22:59.190 --> 00:23:01.750 - But could we be supporting AI companies in 614 00:23:01.750 --> 00:23:03.210 - making sure that the content 615 00:23:03.509 --> 00:23:04.009 - that 616 00:23:04.710 --> 00:23:06.569 - they use and that they produce 617 00:23:06.950 --> 00:23:09.829 - from their large language models or or small 618 00:23:09.829 --> 00:23:10.605 - language models 619 00:23:11.565 --> 00:23:13.664 - is also validated and filtrated. 620 00:23:14.044 --> 00:23:16.365 - So I think that that, you know, we 621 00:23:16.365 --> 00:23:18.384 - are people that know under we're organizations. 622 00:23:18.845 --> 00:23:19.904 - We're people. We're 623 00:23:20.444 --> 00:23:22.684 - an industry that knows how to produce trusted 624 00:23:22.684 --> 00:23:24.845 - content, and that's what the AI companies need 625 00:23:24.845 --> 00:23:26.210 - right now. So, Antonia, 626 00:23:26.669 --> 00:23:29.630 - many other things in scholarly publishing have changed 627 00:23:29.630 --> 00:23:30.130 - dramatically 628 00:23:30.509 --> 00:23:32.049 - over the past thirty years. 629 00:23:32.909 --> 00:23:35.630 - The demise of paper, for example, and the 630 00:23:35.630 --> 00:23:39.150 - move to open access publishing are two very 631 00:23:39.150 --> 00:23:39.650 - important 632 00:23:40.325 --> 00:23:41.704 - examples of change. 633 00:23:42.404 --> 00:23:44.984 - How has this changed the role of publishers, 634 00:23:45.365 --> 00:23:47.765 - and what further changes can we expect in 635 00:23:47.765 --> 00:23:48.345 - the future? 636 00:23:49.125 --> 00:23:52.005 - Yeah. You're right. The research land landscape has 637 00:23:52.005 --> 00:23:54.549 - changed dramatically over the last thirty years, and 638 00:23:54.549 --> 00:23:56.569 - that's thirty years I've been working in 639 00:23:56.950 --> 00:23:58.009 - scholarly publishing. 640 00:23:58.390 --> 00:24:00.569 - But I think the fundamental purpose of publishers 641 00:24:00.869 --> 00:24:04.169 - hasn't changed, and that's really around helping researchers 642 00:24:04.390 --> 00:24:05.929 - communicate trusted knowledge 643 00:24:06.355 --> 00:24:08.214 - and then helping others to 644 00:24:08.674 --> 00:24:10.615 - discover and then build on that knowledge 645 00:24:11.234 --> 00:24:13.875 - ultimately for the benefit of society. So, you 646 00:24:13.875 --> 00:24:15.474 - know, I think I think the world around 647 00:24:15.474 --> 00:24:17.554 - us changed, but the the the purpose of 648 00:24:17.554 --> 00:24:18.774 - publishing hasn't. 649 00:24:19.930 --> 00:24:21.869 - But one of the challenges for publishers 650 00:24:22.170 --> 00:24:22.910 - is that 651 00:24:23.289 --> 00:24:25.630 - some of our most important contributions 652 00:24:26.890 --> 00:24:28.190 - are behind the scenes. 653 00:24:28.730 --> 00:24:31.450 - So researchers rightly focus on the article in 654 00:24:31.450 --> 00:24:33.554 - front of them, but there's a huge amount 655 00:24:33.554 --> 00:24:36.694 - of infrastructure and quality assurance and stewardship 656 00:24:37.714 --> 00:24:40.294 - that sat sat behind that article. 657 00:24:41.234 --> 00:24:42.775 - And that's something publishers 658 00:24:43.394 --> 00:24:45.015 - haven't done a good job of 659 00:24:45.394 --> 00:24:45.894 - describing, 660 00:24:46.674 --> 00:24:48.329 - that that value that we add. 661 00:24:49.049 --> 00:24:51.769 - And as the research ecosystem becomes more complex 662 00:24:51.769 --> 00:24:53.309 - and and AI driven, 663 00:24:53.769 --> 00:24:56.750 - those functions become more important, not less. 664 00:24:57.289 --> 00:24:59.609 - So in the past, we've typically been judged 665 00:24:59.609 --> 00:25:00.109 - on, 666 00:25:01.130 --> 00:25:03.789 - you know, our ability to distribute the content. 667 00:25:03.944 --> 00:25:06.264 - You know, electronic made it more easy to 668 00:25:06.264 --> 00:25:08.345 - disc to discover it and find it all 669 00:25:08.345 --> 00:25:10.984 - around the world. Open access then made it 670 00:25:10.984 --> 00:25:13.384 - even easier, you know, barriers to be able 671 00:25:13.384 --> 00:25:15.224 - to find it. But in the future, I 672 00:25:15.224 --> 00:25:16.764 - think we'll we'll be judged 673 00:25:17.080 --> 00:25:20.039 - much more by our ability, again, to help 674 00:25:20.039 --> 00:25:22.140 - people identify the trusted content, 675 00:25:22.519 --> 00:25:24.519 - sort the chaff from the is that the 676 00:25:24.519 --> 00:25:26.119 - right expression? Sort the chaff from the chaff. 677 00:25:26.119 --> 00:25:28.279 - Weak from the chaff. Weak from the chaff, 678 00:25:28.279 --> 00:25:29.820 - I think. Thank you. Yeah. 679 00:25:30.784 --> 00:25:32.565 - So I think, you know, again, 680 00:25:33.265 --> 00:25:36.244 - research has been consumed through journal articles, 681 00:25:37.585 --> 00:25:40.464 - and we've moved those journal articles from print 682 00:25:40.464 --> 00:25:42.544 - to digital and from behind a paywall to 683 00:25:42.544 --> 00:25:43.605 - now freely available. 684 00:25:44.130 --> 00:25:46.309 - And I think the future is now about 685 00:25:46.529 --> 00:25:47.750 - making sure that, 686 00:25:49.410 --> 00:25:51.730 - we can we can put that information into 687 00:25:51.730 --> 00:25:54.609 - the workflows where people need it. And that 688 00:25:54.609 --> 00:25:56.529 - means our role as publishers is going to 689 00:25:56.529 --> 00:25:59.835 - evolve again to look different. But the purpose 690 00:25:59.835 --> 00:26:01.055 - will remain the same. 691 00:26:01.994 --> 00:26:03.055 - But that that 692 00:26:03.674 --> 00:26:05.295 - that sort of custodian 693 00:26:06.154 --> 00:26:09.835 - of trusted content, not just content, but trusted 694 00:26:09.835 --> 00:26:11.615 - content, I think becomes the imperative. 695 00:26:12.349 --> 00:26:14.690 - And that that that whole sort of custodian 696 00:26:14.910 --> 00:26:16.990 - role is an interesting one because, you know, 697 00:26:16.990 --> 00:26:17.970 - thirty years ago, 698 00:26:18.349 --> 00:26:21.109 - that role was still being played by libraries, 699 00:26:21.109 --> 00:26:23.730 - you know, big university libraries. 700 00:26:24.109 --> 00:26:26.154 - They had their stacks and stacks of paper 701 00:26:26.154 --> 00:26:29.115 - journals, and that's where the information was kept 702 00:26:29.115 --> 00:26:29.615 - safely, 703 00:26:30.474 --> 00:26:32.494 - and could be accessed by people. 704 00:26:32.795 --> 00:26:34.414 - But now, I mean, 705 00:26:35.115 --> 00:26:36.494 - is it one of the unintended 706 00:26:36.795 --> 00:26:37.295 - consequences 707 00:26:37.914 --> 00:26:39.994 - of getting rid of paper for the most 708 00:26:39.994 --> 00:26:42.309 - part? Is that now it's down to 709 00:26:42.769 --> 00:26:46.309 - academic publishers to to maintain that record. It's, 710 00:26:46.450 --> 00:26:48.210 - you know, sort of is it on you 711 00:26:48.210 --> 00:26:50.150 - now? Or or do you or is 712 00:26:50.609 --> 00:26:53.490 - do you still send paper journals to to 713 00:26:53.490 --> 00:26:55.410 - libraries? You know, it's funny. I work for 714 00:26:55.410 --> 00:26:57.664 - IOP publishing, and I don't know. I I 715 00:26:57.664 --> 00:27:00.065 - assume that stopped, but maybe It's pretty much 716 00:27:00.065 --> 00:27:01.444 - stopped. There are a few, 717 00:27:01.865 --> 00:27:02.565 - a few 718 00:27:02.865 --> 00:27:05.105 - rare libraries around the world that still require 719 00:27:05.105 --> 00:27:06.325 - print, but our digital, 720 00:27:07.184 --> 00:27:09.924 - our our print footprint is so small now, 721 00:27:11.200 --> 00:27:13.679 - that most but but there is still a 722 00:27:13.679 --> 00:27:16.240 - a way to archive digital versions, and that's 723 00:27:16.240 --> 00:27:18.000 - what we need to continue to do. So 724 00:27:18.000 --> 00:27:20.079 - we submit all of our content to clocks 725 00:27:20.079 --> 00:27:21.380 - as one one archiving 726 00:27:21.919 --> 00:27:24.154 - solution, for example. But I think the role 727 00:27:24.154 --> 00:27:26.335 - of librarians, again, just like publishers, 728 00:27:26.954 --> 00:27:29.375 - is going to shift. I mean, librarians ultimately 729 00:27:29.434 --> 00:27:30.815 - have to enable 730 00:27:31.674 --> 00:27:35.194 - access to research for their patrons so that 731 00:27:35.194 --> 00:27:36.875 - they can do more research or that they 732 00:27:36.875 --> 00:27:39.240 - can learn to be a researcher if they're 733 00:27:39.240 --> 00:27:40.220 - students. So 734 00:27:40.599 --> 00:27:42.519 - the the the sort of gateway role of 735 00:27:42.519 --> 00:27:45.720 - a librarian between all the content that is 736 00:27:45.720 --> 00:27:48.119 - available out there in the world and, you 737 00:27:48.119 --> 00:27:49.019 - know, the discoverability 738 00:27:49.480 --> 00:27:51.740 - for their faculty and students 739 00:27:52.484 --> 00:27:54.505 - remains. But, again, increasingly, 740 00:27:55.525 --> 00:27:56.424 - that will be, 741 00:27:57.444 --> 00:27:58.585 - mediated through 742 00:27:59.204 --> 00:27:59.704 - machines. 743 00:28:00.805 --> 00:28:02.724 - You you know, you see examples now of 744 00:28:02.724 --> 00:28:05.400 - researchers, particularly in physics, because they're pretty, 745 00:28:06.099 --> 00:28:06.599 - technology 746 00:28:06.900 --> 00:28:08.200 - minded and and, 747 00:28:08.580 --> 00:28:10.660 - you know, innovative in their in their thinking. 748 00:28:10.660 --> 00:28:14.259 - And you're seeing physicists using tools like Claude's 749 00:28:14.259 --> 00:28:14.759 - science 750 00:28:15.539 --> 00:28:17.940 - to basically do all their research work within 751 00:28:17.940 --> 00:28:19.625 - this one environment. They don't wanna have to 752 00:28:19.625 --> 00:28:21.384 - come out of that to go and check 753 00:28:21.384 --> 00:28:23.785 - a reference source or to go and find 754 00:28:23.785 --> 00:28:24.285 - some 755 00:28:24.585 --> 00:28:26.904 - some more information about a particular topic. They 756 00:28:26.904 --> 00:28:28.265 - want to be able to do that within 757 00:28:28.265 --> 00:28:30.265 - the same environment that they're using to do 758 00:28:30.265 --> 00:28:30.924 - their research. 759 00:28:31.625 --> 00:28:34.365 - So the onus is on publishers and librarians 760 00:28:34.664 --> 00:28:35.325 - to create 761 00:28:35.920 --> 00:28:37.140 - a seamless access 762 00:28:37.440 --> 00:28:37.940 - between 763 00:28:38.559 --> 00:28:40.880 - the the AI tooling that the researchers are 764 00:28:40.880 --> 00:28:43.359 - using or the students are using and the 765 00:28:43.359 --> 00:28:46.559 - content that we continue to produce through the 766 00:28:46.559 --> 00:28:49.779 - through the research process. So librarians will increasingly, 767 00:28:51.865 --> 00:28:55.785 - be engaged in supporting their, patrons, whether those 768 00:28:55.785 --> 00:28:57.244 - are faculty or students, 769 00:28:57.865 --> 00:29:00.444 - in interacting with the scientific literature 770 00:29:01.065 --> 00:29:04.025 - via things like MCPs, which are called model 771 00:29:04.025 --> 00:29:07.539 - context protocols or APIs, which people will have 772 00:29:07.539 --> 00:29:10.179 - heard of. But these are sort of gateways, 773 00:29:10.179 --> 00:29:11.559 - if you like, between 774 00:29:12.339 --> 00:29:15.399 - an AI tool and the the established literature. 775 00:29:15.859 --> 00:29:18.375 - And so I think, you know, librarians are 776 00:29:19.335 --> 00:29:20.954 - similarly gonna have to evolve 777 00:29:21.255 --> 00:29:22.234 - what they do, 778 00:29:22.615 --> 00:29:25.095 - but their purpose will remain the same just 779 00:29:25.095 --> 00:29:26.315 - as publishers' purposes 780 00:29:26.694 --> 00:29:28.234 - purpose will remain the same. 781 00:29:28.535 --> 00:29:29.674 - So I think it's a technology 782 00:29:30.055 --> 00:29:30.555 - evolution, 783 00:29:31.894 --> 00:29:32.714 - rather than 784 00:29:33.255 --> 00:29:34.640 - a kind of disintermediation. 785 00:29:36.059 --> 00:29:37.580 - Well, I'm glad we we got to speak 786 00:29:37.580 --> 00:29:39.900 - about librarians because I know that here at 787 00:29:39.900 --> 00:29:41.580 - IOP publishing, we do have a lot of 788 00:29:41.580 --> 00:29:42.080 - interaction 789 00:29:42.539 --> 00:29:45.500 - with librarians, but often they're left out of 790 00:29:45.500 --> 00:29:48.240 - this discussion when it comes to scholarly publishing. 791 00:29:48.734 --> 00:29:51.375 - So you've spent your career helping people connect 792 00:29:51.375 --> 00:29:53.474 - with trusted research and knowledge. 793 00:29:53.855 --> 00:29:56.174 - What words of encouragement would you give to 794 00:29:56.174 --> 00:29:58.434 - someone who is interested in embarking 795 00:29:58.734 --> 00:30:01.795 - on a similar career in scholarly publishing? 796 00:30:02.669 --> 00:30:03.950 - Well, I think it's one of the most 797 00:30:03.950 --> 00:30:05.009 - exciting times 798 00:30:05.950 --> 00:30:08.190 - to consider a career in scholarly publishing, so 799 00:30:08.190 --> 00:30:11.250 - I would absolutely encourage everyone to do so. 800 00:30:11.309 --> 00:30:13.069 - I think I've been saying that for thirty 801 00:30:13.069 --> 00:30:15.704 - years. But, genuinely, I think the 802 00:30:16.184 --> 00:30:17.644 - the pace of change, 803 00:30:18.424 --> 00:30:18.924 - the 804 00:30:19.224 --> 00:30:19.724 - uncertainty 805 00:30:20.184 --> 00:30:22.284 - that we face now is is unprecedented. 806 00:30:23.144 --> 00:30:25.065 - And for some people, that that might be 807 00:30:25.065 --> 00:30:25.565 - frightening. 808 00:30:26.105 --> 00:30:28.845 - But if you're somebody that really enjoys, 809 00:30:30.419 --> 00:30:30.919 - learning, 810 00:30:31.539 --> 00:30:33.480 - and who's curious to understand 811 00:30:34.179 --> 00:30:36.980 - things that are changing and intersections that are 812 00:30:36.980 --> 00:30:40.759 - shifting between research people and emerging technologies, then 813 00:30:40.980 --> 00:30:42.200 - then it's a fantastically 814 00:30:43.095 --> 00:30:43.595 - interesting 815 00:30:44.054 --> 00:30:44.554 - profession 816 00:30:44.934 --> 00:30:47.335 - to to come into. I mean, what attracted 817 00:30:47.335 --> 00:30:48.394 - me to the sector 818 00:30:49.015 --> 00:30:51.654 - thirty plus years ago has has kept me 819 00:30:51.654 --> 00:30:53.654 - engaged ever since, and that is that sort 820 00:30:53.654 --> 00:30:54.154 - of 821 00:30:54.535 --> 00:30:55.674 - intersection between 822 00:30:56.210 --> 00:30:57.349 - science, technology, 823 00:30:58.369 --> 00:31:00.390 - creativity. I mean, a lot of people think 824 00:31:00.930 --> 00:31:04.390 - creativity in publishing is, you know, magazine publishing 825 00:31:05.089 --> 00:31:07.009 - or publishing novels or whatever, but there's a 826 00:31:07.009 --> 00:31:09.269 - lot of creativity in scientific publishing. 827 00:31:10.115 --> 00:31:12.275 - And, of course, the communication piece. We have 828 00:31:12.275 --> 00:31:15.255 - to be expert in communication, in scientific communication. 829 00:31:15.875 --> 00:31:17.975 - So I think, you know, it's pretty motivational 830 00:31:18.115 --> 00:31:20.215 - to think that every day you're helping researchers 831 00:31:21.075 --> 00:31:23.235 - get their research out into the world to 832 00:31:23.235 --> 00:31:26.190 - be built upon and and hopefully to improve, 833 00:31:27.049 --> 00:31:30.109 - you know, improve society through through saving, 834 00:31:30.890 --> 00:31:33.849 - saving lives or, you know, addressing some of 835 00:31:33.849 --> 00:31:36.734 - the massive societal challenges that we we face. 836 00:31:37.054 --> 00:31:38.994 - So I think it's it's got a purpose, 837 00:31:39.214 --> 00:31:40.734 - which has always been a big thing for 838 00:31:40.734 --> 00:31:41.234 - me. 839 00:31:41.855 --> 00:31:43.634 - It's got the intellectual stimulation, 840 00:31:45.054 --> 00:31:47.474 - and it's got the kind of fast changing, 841 00:31:48.494 --> 00:31:49.875 - you know, constant reinvention 842 00:31:50.255 --> 00:31:51.855 - going for it. And so you put those 843 00:31:51.855 --> 00:31:54.109 - things together, and it's pretty compelling. 844 00:31:54.490 --> 00:31:55.950 - Yeah. I mean, especially, 845 00:31:56.490 --> 00:31:58.990 - I suppose, now that, you know, publishing 846 00:31:59.289 --> 00:32:02.170 - is especially scholarly publishing, you're right at the 847 00:32:02.170 --> 00:32:03.309 - edge of AI. 848 00:32:04.009 --> 00:32:04.750 - You know, your 849 00:32:05.305 --> 00:32:08.285 - your information is crucial for the success of 850 00:32:08.904 --> 00:32:11.625 - AI. You know, that interaction must be must 851 00:32:11.625 --> 00:32:13.244 - be fascinating, I would imagine. 852 00:32:13.865 --> 00:32:15.865 - I I, you know, I was starting out 853 00:32:15.865 --> 00:32:18.125 - on this this new strategy work, 854 00:32:18.789 --> 00:32:21.289 - having, you know, been working under our previous 855 00:32:21.829 --> 00:32:24.710 - predominantly open access strategy that we had when 856 00:32:24.710 --> 00:32:26.250 - I became CEO. And, 857 00:32:26.630 --> 00:32:28.390 - you know, it it took me a while 858 00:32:28.390 --> 00:32:28.970 - to think, 859 00:32:30.070 --> 00:32:32.170 - where is the nub of this? And it 860 00:32:32.470 --> 00:32:33.529 - it is absolutely 861 00:32:34.765 --> 00:32:37.724 - re AI is reshaping everything we do. And 862 00:32:37.724 --> 00:32:38.704 - that's terrifying, 863 00:32:39.164 --> 00:32:41.804 - but it's also incredibly exciting and it opens 864 00:32:41.804 --> 00:32:43.025 - up huge opportunities. 865 00:32:43.484 --> 00:32:44.224 - And particularly, 866 00:32:44.684 --> 00:32:47.025 - I think, for a society publisher like 867 00:32:47.450 --> 00:32:49.950 - IAP Publishing because that trust 868 00:32:50.569 --> 00:32:52.490 - we we we have a reputation and we 869 00:32:52.490 --> 00:32:54.349 - have trust in the market that's 870 00:32:54.730 --> 00:32:57.069 - pretty high compared to many commercial 871 00:32:57.450 --> 00:32:57.950 - publishers. 872 00:32:58.329 --> 00:33:00.730 - We're also, you know, deeply rooted to the 873 00:33:00.730 --> 00:33:01.230 - communities 874 00:33:01.609 --> 00:33:03.884 - that we serve, broader than physics, but all 875 00:33:03.884 --> 00:33:05.904 - within the kind of physical science area. 876 00:33:06.524 --> 00:33:08.464 - So that confers quite a lot of advantage 877 00:33:08.605 --> 00:33:11.164 - when you think about how we navigate through 878 00:33:11.164 --> 00:33:13.644 - and still remain relevant in this sort of 879 00:33:13.644 --> 00:33:14.944 - AI era. So 880 00:33:15.349 --> 00:33:17.529 - I I think it presents us with a 881 00:33:17.750 --> 00:33:18.970 - a whole new sort 882 00:33:20.390 --> 00:33:21.049 - of toolkit 883 00:33:21.589 --> 00:33:23.830 - to be more relevant and more successful in 884 00:33:23.830 --> 00:33:24.490 - the future. 885 00:33:25.430 --> 00:33:27.109 - Well, that's great. That's a great place to 886 00:33:27.109 --> 00:33:27.609 - end. 887 00:33:28.230 --> 00:33:30.490 - Thanks for coming on the podcast, Antonia. 888 00:33:30.934 --> 00:33:32.775 - Thank you very much, Hamish. I mean, one 889 00:33:32.775 --> 00:33:34.214 - thing I would just say to end is 890 00:33:34.214 --> 00:33:37.194 - that, you know, we haven't talked about geopolitical 891 00:33:37.335 --> 00:33:37.835 - fragmentation 892 00:33:38.214 --> 00:33:38.714 - or 893 00:33:39.095 --> 00:33:40.315 - scientific nationalism, 894 00:33:40.855 --> 00:33:41.355 - and 895 00:33:41.734 --> 00:33:43.575 - and that's huge in and of itself. We 896 00:33:43.575 --> 00:33:45.994 - could have had another podcast all about that. 897 00:33:46.214 --> 00:33:49.240 - But for me, whatever the changes are ahead, 898 00:33:49.779 --> 00:33:52.440 - that trusted research, international collaboration, 899 00:33:53.140 --> 00:33:55.460 - and human expertise are gonna continue to be 900 00:33:55.460 --> 00:33:56.119 - the foundations 901 00:33:56.420 --> 00:33:59.220 - on which scientific progress is built. So thank 902 00:33:59.220 --> 00:33:59.720 - you. 903 00:34:07.525 --> 00:34:09.385 - That was Antonia Seymour, 904 00:34:09.765 --> 00:34:12.344 - chief executive of IOP Publishing. 905 00:34:13.139 --> 00:34:16.179 - Thanks to Antonia for her insights into the 906 00:34:16.179 --> 00:34:17.960 - challenges and opportunities 907 00:34:18.579 --> 00:34:20.039 - in scholarly publishing. 908 00:34:20.980 --> 00:34:23.059 - I'm afraid that's all the time we have 909 00:34:23.059 --> 00:34:24.359 - for this week's podcast. 910 00:34:24.844 --> 00:34:28.204 - I'm Hamish Johnston, and our producer is Fred 911 00:34:28.204 --> 00:34:28.704 - Iles. 912 00:34:29.164 --> 00:34:31.664 - The music in Physics World's podcasts 913 00:34:31.964 --> 00:34:34.605 - is called one three seven, and it was 914 00:34:34.605 --> 00:34:35.984 - composed and performed 915 00:34:36.284 --> 00:34:37.264 - by the physicist 916 00:34:37.885 --> 00:34:38.944 - Philip Moriarty. 917 00:34:40.070 --> 00:34:41.929 - We'll be back again next week.