1 00:00:12,880 --> 00:00:14,870 Welcome to studying law Around the world. 2 00:00:15,020 --> 00:00:17,870 I'm Claudio Claus. In each episode, I talk with 3 00:00:17,880 --> 00:00:20,850 lawyers, law students, and professors from different parts 4 00:00:20,860 --> 00:00:24,730 of the world to talk about legal education, careers, and what the 5 00:00:24,740 --> 00:00:26,410 profession looks like in real life. 6 00:00:26,420 --> 00:00:29,470 We talk about the hard parts, the surprises and the decisions. 7 00:00:29,480 --> 00:00:32,270 That shaped their paths. Whether you're planning to study 8 00:00:32,280 --> 00:00:35,520 abroad, thinking about working in another legal system, or just 9 00:00:35,530 --> 00:00:39,630 curious about how law works around the world, this podcast 10 00:00:39,640 --> 00:00:45,160 is for you. If you're an internationally 11 00:00:45,170 --> 00:00:48,480 trained lawyer or law student working in or trying to break 12 00:00:48,490 --> 00:00:52,880 into an English speaking common law system, you know that legal 13 00:00:52,890 --> 00:00:56,300 writing is one of the biggest challenges and one of the most 14 00:00:56,310 --> 00:01:00,260 important skills to master. That's where Grammatica comes 15 00:01:00,270 --> 00:01:02,460 in. I'm Rebecca Lockwood, the 16 00:01:02,470 --> 00:01:05,310 lawyer, mediator, educator and founder of Grammatica 17 00:01:05,319 --> 00:01:08,780 International. We help lawyers and law students 18 00:01:08,790 --> 00:01:13,320 to write sharper, speak clearer, and land more interviews with on 19 00:01:13,330 --> 00:01:17,280 demand online courses, live workshops and one-on-one 20 00:01:17,290 --> 00:01:20,060 consulting designed just for legal professionals. 21 00:01:20,420 --> 00:01:23,510 I have worked with so many lawyers like you who are 22 00:01:23,520 --> 00:01:27,040 exceptionally smart, qualified and capable, but need some 23 00:01:27,050 --> 00:01:29,980 guidance to really stand out in a competitive legal market. 24 00:01:30,150 --> 00:01:32,800 And I get it myself, having lived and worked in five 25 00:01:32,810 --> 00:01:35,950 countries during my 17 years in law and education. 26 00:01:36,630 --> 00:01:39,080 Whether you're looking to sharpen your writing skills 27 00:01:39,090 --> 00:01:42,960 generally, refresh your CD, or draft an application for a law 28 00:01:42,970 --> 00:01:47,970 program, Grammatica can help. Visit Grammatico to get started. 29 00:01:52,500 --> 00:01:55,130 Today we're having two guests together here. 30 00:01:55,180 --> 00:02:00,090 So we have Michael Cooper and also Spencer Nayer with us. 31 00:02:00,180 --> 00:02:03,430 They've been all over the news in the legal world because of 32 00:02:03,440 --> 00:02:06,370 something very interesting that they did at their law school 33 00:02:06,380 --> 00:02:09,669 journal, so at the Texas A&M Journal of Property Law. 34 00:02:09,780 --> 00:02:13,360 And that was coming up with a whole edition of the journal, 35 00:02:13,370 --> 00:02:21,060 whole issue, which was very much all about AI, but also with AI 36 00:02:21,070 --> 00:02:23,230 doing that. And we'll be talking about how 37 00:02:23,240 --> 00:02:27,100 that whole process went. Thank you so much both of you to 38 00:02:27,150 --> 00:02:30,380 to join us today and also wanted to ask you to introduce yourself 39 00:02:30,390 --> 00:02:32,660 before we go on into the questions of the show. 40 00:02:32,730 --> 00:02:36,280 Yeah, Thanks for having us. My name is Michael Cooper. 41 00:02:36,290 --> 00:02:40,700 I am a recent graduate of the Texas A&M School of Law. 42 00:02:40,710 --> 00:02:43,220 I'll be practicing in Dallas this fall. 43 00:02:43,310 --> 00:02:46,440 And I was the managing editor of the journal Property Law. 44 00:02:46,510 --> 00:02:49,040 Yeah, just echoing it. Thanks for having us. 45 00:02:49,050 --> 00:02:51,960 Very excited to be here. My name is Spencer Nyer. 46 00:02:51,970 --> 00:02:54,250 I am similarly. A graduate. 47 00:02:54,260 --> 00:02:57,420 Recent graduate. I'll be practicing patent law 48 00:02:57,430 --> 00:03:00,920 here in a few months, hopefully after the bar and. 49 00:03:01,070 --> 00:03:03,390 I was the. Editor in chief of the Journal 50 00:03:03,400 --> 00:03:04,370 of. Property Law. 51 00:03:04,720 --> 00:03:06,840 During the time that this issue came out. 52 00:03:06,850 --> 00:03:09,650 Fantastic. Before we jump into the issue 53 00:03:09,660 --> 00:03:12,360 itself, now that we know a little bit more about you, I 54 00:03:12,370 --> 00:03:15,680 also wanted to ask you just briefly about your journey to 55 00:03:15,690 --> 00:03:18,190 law school. So was there something specific 56 00:03:18,200 --> 00:03:20,360 that sparked your interesting law? 57 00:03:20,400 --> 00:03:24,490 And then also did that path kind of automatically took you into 58 00:03:24,500 --> 00:03:28,370 being involved in the in the law journal or you know, that kind 59 00:03:28,380 --> 00:03:30,650 of happened along the way? I'd love to hear your thoughts. 60 00:03:30,700 --> 00:03:35,010 Yeah, so my path to law school was atypical. 61 00:03:35,020 --> 00:03:38,550 I wasn't someone who always knew that law was going to be my 62 00:03:38,560 --> 00:03:40,400 path. Originally I thought I was going 63 00:03:40,410 --> 00:03:44,110 to be an engineer and so I went to school for electrical 64 00:03:44,120 --> 00:03:46,530 engineering and somewhere along the. 65 00:03:46,540 --> 00:03:50,110 Way I was, I had. Graduated and started looking at 66 00:03:50,120 --> 00:03:55,390 career field and somebody made some offhand comment about 67 00:03:55,440 --> 00:03:59,790 patent law and then I just kind. Of researched the career the 68 00:03:59,800 --> 00:04:02,370 field would I be. Doing and things just kind of 69 00:04:02,380 --> 00:04:06,090 started falling into place. It was accepted to and and you 70 00:04:06,100 --> 00:04:09,290 know, now graduated. I kind of always knew. 71 00:04:09,300 --> 00:04:11,550 That. With patent laws, my interest 72 00:04:11,560 --> 00:04:14,270 property was kind of the path I was going to take, intellectual 73 00:04:14,280 --> 00:04:16,790 property. And so for me, the Journal of 74 00:04:16,800 --> 00:04:19,510 Property Law made a lot of sense because it really aligned with a 75 00:04:19,519 --> 00:04:21,010 lot of the things that I wanted to do. 76 00:04:21,430 --> 00:04:25,730 Yeah, and my, my path to law school was, if not atypical, 77 00:04:25,740 --> 00:04:28,470 probably like the fairy tale way that people decide they want to 78 00:04:28,480 --> 00:04:31,650 go to law school. I was four or five years old and 79 00:04:31,660 --> 00:04:35,080 I was watching West Wing and Boston Legal and, and all of 80 00:04:35,090 --> 00:04:36,890 these shows. And there was just something 81 00:04:36,900 --> 00:04:40,290 unique about the way lawyers talked and carried themselves. 82 00:04:40,300 --> 00:04:43,470 And I'm not from a family of lawyers, but I just kind of 83 00:04:43,480 --> 00:04:47,490 attached to that idea and image of, of a professional who, who 84 00:04:47,500 --> 00:04:51,470 gives counsel, who's the smart person in the room, the reliable 85 00:04:51,480 --> 00:04:53,450 person in the room. And so that's always what I 86 00:04:53,460 --> 00:04:55,530 wanted to be since I was four or five years old. 87 00:04:55,540 --> 00:05:01,150 I graduated undergrad in 2017, moved to DC, worked in higher Ed 88 00:05:01,160 --> 00:05:05,390 and science policy with kind of a, a particular focus in 89 00:05:05,710 --> 00:05:08,760 automation and enabling automation research. 90 00:05:08,770 --> 00:05:13,890 And then I moved back to Texas when I got married and worked 91 00:05:13,900 --> 00:05:16,740 for the ledge for a couple years and then eventually was able to 92 00:05:16,750 --> 00:05:19,750 matriculate and got, except when I got accepted into A&M. 93 00:05:20,230 --> 00:05:24,100 And so the Journal of property law, I knew I wanted to keep 94 00:05:24,110 --> 00:05:26,650 clerkships open as an opportunity. 95 00:05:26,710 --> 00:05:29,580 And so I knew I needed to be on a publication at A&M. 96 00:05:29,590 --> 00:05:31,700 We have two, we have the journal property Law and we have the law 97 00:05:31,710 --> 00:05:33,460 review. And I got accepted on to the 98 00:05:33,470 --> 00:05:36,420 Journal of Property law. And I, I do find property law 99 00:05:36,430 --> 00:05:38,980 incredibly interesting. And so it was, it was a nice 100 00:05:38,990 --> 00:05:42,580 intellectual pursuit and the opportunity eventually to, to be 101 00:05:42,590 --> 00:05:44,840 on the executive board and kind of steer that body of 102 00:05:44,850 --> 00:05:48,520 scholarship as well. And so I think there were a lot 103 00:05:48,530 --> 00:05:51,570 of different forces that pulled me into the direction of working 104 00:05:51,580 --> 00:05:54,170 on this particular drone. Amazing, That's awesome. 105 00:05:54,240 --> 00:05:57,890 And I wanted to hear all about, you know, going into the this 106 00:05:57,900 --> 00:06:01,200 specific publication. So you both mentioned about, you 107 00:06:01,210 --> 00:06:04,010 know, kind of getting to practice, finishing up law 108 00:06:04,020 --> 00:06:07,290 school now and going to take the bar and all of that. 109 00:06:07,340 --> 00:06:10,410 So I'd love to hear, you know, on the start on on kind of 110 00:06:10,420 --> 00:06:13,390 looking ahead, what are some ways that you see this is 111 00:06:13,400 --> 00:06:17,550 specific publication, you know, changing the experience of of 112 00:06:17,560 --> 00:06:21,360 future law journals all around America, but maybe all around 113 00:06:21,370 --> 00:06:25,670 the world and influencing some kind of the the direction of how 114 00:06:25,760 --> 00:06:27,430 things are going to get done from now on. 115 00:06:27,500 --> 00:06:31,470 Definitely, I think like as a, in my role as managing editor, 116 00:06:31,540 --> 00:06:35,080 my job is was to curate the scholarship for the journal, 117 00:06:35,090 --> 00:06:38,060 select that and approve it for final publication. 118 00:06:38,290 --> 00:06:42,400 And I think that in that role, seeing products that are 119 00:06:42,410 --> 00:06:46,760 produced in conjunction with artificial intelligence is going 120 00:06:46,770 --> 00:06:49,180 to become part of the new normal. 121 00:06:49,230 --> 00:06:53,300 This is something that all facets of academia are currently 122 00:06:53,350 --> 00:06:55,680 observing and dealing with, whether it's publication in 123 00:06:55,690 --> 00:06:57,920 books, whether it's publication in peer reviewed medical 124 00:06:57,930 --> 00:07:01,140 journals or other kinds of social science and things like 125 00:07:01,150 --> 00:07:03,300 that. Artificial intelligence is part 126 00:07:03,350 --> 00:07:07,480 of the body scholarly. And so because of that, my hope 127 00:07:07,490 --> 00:07:11,650 is that our foray that we explore not only in our forward 128 00:07:11,660 --> 00:07:15,890 that we wrote about the works, but also looking at the works 129 00:07:15,900 --> 00:07:18,860 that themselves published provides a little bit of a 130 00:07:18,870 --> 00:07:24,640 grounding experience for law review staff to think about the 131 00:07:24,650 --> 00:07:27,500 role that AI can play in this process. 132 00:07:27,510 --> 00:07:30,260 To hopefully kind of, we talked about it in the forward, but 133 00:07:30,270 --> 00:07:33,410 kind of get rid of the the scarlet AI kind of thinking 134 00:07:33,420 --> 00:07:36,280 about like Nathaniel Hawthorne there and, and trying to 135 00:07:36,290 --> 00:07:39,070 eradicate more of the stigma so that they can have a more 136 00:07:39,080 --> 00:07:42,880 constructive dialogue about the product rather than the process 137 00:07:42,930 --> 00:07:46,740 to be able to publish works that meaningfully contribute to the 138 00:07:46,750 --> 00:07:48,100 scholarship. Yeah. 139 00:07:48,110 --> 00:07:51,000 Absolutely. And just to echo, any new 140 00:07:51,010 --> 00:07:54,660 technology has benefits and drawbacks. 141 00:07:54,990 --> 00:07:59,480 And as AI continues to roll out and become more and more 142 00:07:59,490 --> 00:08:03,370 prevalent, something that's kind of already clear in scholarship 143 00:08:03,380 --> 00:08:04,720 in general. Across fields is. 144 00:08:04,730 --> 00:08:08,620 That AI is being. Used to write papers that. 145 00:08:08,630 --> 00:08:12,850 Are being published in journals and things around the country, 146 00:08:12,860 --> 00:08:18,830 around the world, and so this is already happening the problem is 147 00:08:18,840 --> 00:08:24,950 that there's no overarching scheme or thought on what kind 148 00:08:24,960 --> 00:08:26,530 of. Disclosure needs to be used. 149 00:08:26,540 --> 00:08:29,900 What kind of extra editing or extra processing or extra 150 00:08:29,910 --> 00:08:33,630 thinking needs to be done when these things are being 151 00:08:33,640 --> 00:08:36,970 published? And so you have these journals, 152 00:08:36,980 --> 00:08:41,350 publications that are publishing AI articles with absolutely no 153 00:08:41,360 --> 00:08:44,760 idea that it's AI. And that's really kind of what 154 00:08:44,770 --> 00:08:47,520 we were trying to address in the forward and. 155 00:08:47,590 --> 00:08:49,980 With this issue is the fact that. 156 00:08:50,030 --> 00:08:53,580 This is happening now and this is going to continue to happen 157 00:08:53,770 --> 00:08:57,820 more and more frequently and so there needs to be some sort of 158 00:08:57,830 --> 00:09:00,190 scheme. For what is moral? 159 00:09:00,200 --> 00:09:02,120 With the use of AI and what is ethical? 160 00:09:02,130 --> 00:09:04,300 For. Publication of these kinds of 161 00:09:04,310 --> 00:09:08,390 materials. It was quite the bold move to to 162 00:09:08,400 --> 00:09:12,550 publish a full volume that is of AI assisted scholarship. 163 00:09:12,640 --> 00:09:15,970 And I'm wondering here, what was some of the the turning points 164 00:09:15,980 --> 00:09:18,830 that convinced you this was the right step for the journal? 165 00:09:18,880 --> 00:09:23,530 And if initially you found a lot of pushback or did you find more 166 00:09:23,720 --> 00:09:27,590 curiosity within the community, what was what was that like 167 00:09:27,640 --> 00:09:30,170 preparing this? Issue I I think at a high level, 168 00:09:30,220 --> 00:09:32,430 you know, students are basically told when they enter into law 169 00:09:32,440 --> 00:09:36,150 school don't touch AI at all and it's unethical to use it. 170 00:09:36,240 --> 00:09:38,590 And I think that that, particularly in talking to our 171 00:09:38,600 --> 00:09:41,210 advisors, really colours a lot of the way that students 172 00:09:41,260 --> 00:09:45,010 interact with the idea of doing novel things is that they are 173 00:09:45,020 --> 00:09:47,730 presented with these like dos and don'ts when they're being 174 00:09:47,740 --> 00:09:50,560 graded. And so I think a lot of the the 175 00:09:50,570 --> 00:09:53,800 hesitation that we received was from folks who are like, I'm not 176 00:09:53,810 --> 00:09:55,300 allowed to do this in the classroom. 177 00:09:55,310 --> 00:09:57,580 Why would a professor be allowed to do this? 178 00:09:57,650 --> 00:10:01,220 And so I think that was a lot of the initial pushback that we 179 00:10:01,230 --> 00:10:03,020 got. Where were those kind of ethical 180 00:10:03,030 --> 00:10:06,240 concerns? I think the opportunity to 181 00:10:06,590 --> 00:10:08,500 publish. These articles was very 182 00:10:08,510 --> 00:10:10,520 exciting. Like I said, I come from a 183 00:10:10,530 --> 00:10:15,460 technology background and I did some work programming AI systems 184 00:10:15,470 --> 00:10:17,500 during undergrad. And so this is always. 185 00:10:17,510 --> 00:10:21,090 Been kind of an area that I found very fascinating and I. 186 00:10:21,100 --> 00:10:24,480 Think that was a large. Part of the reason because 187 00:10:24,490 --> 00:10:29,780 seeing how far the technology has come just in the. 188 00:10:29,840 --> 00:10:35,070 Four or five years since. I really looking at them, it's 189 00:10:35,080 --> 00:10:38,310 it's pretty incredible. And so you have this opportunity 190 00:10:38,320 --> 00:10:43,210 to look at this new kind of budding technology that is now 191 00:10:43,320 --> 00:10:47,350 being implemented more and more into the mainstream of everyday 192 00:10:47,360 --> 00:10:50,970 life, and you have this opportunity to start talking. 193 00:10:50,980 --> 00:10:53,250 About OK, it's here, how do we? Use it. 194 00:10:53,260 --> 00:10:55,850 How do we benefit from? It what is the? 195 00:10:55,860 --> 00:10:59,610 True benefit of using this technology because if you ask 196 00:10:59,620 --> 00:11:02,230 any person off the streets. Like what is AI? 197 00:11:02,240 --> 00:11:04,110 Why would you use it? It's. 198 00:11:04,340 --> 00:11:07,690 Something along the lines of, well, it can accumulate all this 199 00:11:07,700 --> 00:11:09,310 information and it spits it out faster. 200 00:11:09,320 --> 00:11:13,110 It's more accurate than just a normal Internet search and it 201 00:11:13,120 --> 00:11:16,020 saves you a lot of time and gets you gets the ball rolling on 202 00:11:16,030 --> 00:11:19,840 whatever you're talking about. And so seeing how you can use 203 00:11:19,850 --> 00:11:22,640 that and distill it. Into OK, so. 204 00:11:22,710 --> 00:11:24,250 What I'm hearing is time savings. 205 00:11:24,260 --> 00:11:27,220 What I'm hearing is accuracy. What I'm hearing is the 206 00:11:27,230 --> 00:11:30,400 accumulation of all of this informational resources. 207 00:11:30,410 --> 00:11:32,780 And so how can we take those things and. 208 00:11:32,850 --> 00:11:34,930 Apply it to scholarship. And so that. 209 00:11:34,940 --> 00:11:37,480 Was a really cool opportunity. And then of course, talking 210 00:11:37,490 --> 00:11:41,180 about the. Ethics of the whole situation 211 00:11:41,230 --> 00:11:46,060 to. To coops point, plagiarism and 212 00:11:46,070 --> 00:11:49,370 copying and things like that, from the time you're in, you 213 00:11:49,380 --> 00:11:52,020 know, grade school, that's something that's kind of, you 214 00:11:52,030 --> 00:11:54,920 know, told to every student is that plagiarism is never OK. 215 00:11:54,930 --> 00:11:57,780 This is never OK if you find a source on the Internet. 216 00:11:57,790 --> 00:12:00,550 You can't use it word. Forward and that's not what I. 217 00:12:00,560 --> 00:12:04,890 Does, but AI does. Sort of blur that line of what 218 00:12:04,900 --> 00:12:07,950 is copying what is the accumulation, because there's no 219 00:12:07,960 --> 00:12:11,730 independent thought behind pure AI. 220 00:12:11,780 --> 00:12:17,670 It's really like I fancy Google. Search in in a in a. 221 00:12:17,680 --> 00:12:23,410 So trying to get an output from AI and then. 222 00:12:23,420 --> 00:12:27,380 Use that as a substantial building block for a paper. 223 00:12:27,480 --> 00:12:30,400 Definitely can. Ruffle some feathers for people. 224 00:12:30,410 --> 00:12:34,690 Who have been told for twenty 30-40 years that you know you 225 00:12:34,700 --> 00:12:37,540 can't take something straight off the Internet and and put it 226 00:12:37,600 --> 00:12:39,790 in any substantial. Way into a paper. 227 00:12:39,840 --> 00:12:43,730 Now we're looking at AI that's curating, organizing through 228 00:12:43,780 --> 00:12:46,150 thousands and hundreds of thousands of websites and 229 00:12:46,160 --> 00:12:50,340 articles and then giving you an output that is independent of 230 00:12:50,400 --> 00:12:53,630 any one of those sources, but without independent thought 231 00:12:53,640 --> 00:12:56,470 behind. It is in some ways you could 232 00:12:56,540 --> 00:13:00,870 frame that as like a small copy of millions of different sources 233 00:13:00,880 --> 00:13:03,650 and sighting. And so comparing that to. 234 00:13:03,660 --> 00:13:05,930 What we do as people. That's essentially what we do in 235 00:13:05,940 --> 00:13:07,590 Independent. Thought it's just. 236 00:13:07,660 --> 00:13:11,090 The fact that it's coming from a machine instead of a mind. 237 00:13:12,220 --> 00:13:15,990 Difficult for people to conceptualize. 238 00:13:16,000 --> 00:13:19,360 And accept when those are the lessons you've learned all this 239 00:13:19,370 --> 00:13:21,010 time. And so that was a really 240 00:13:21,020 --> 00:13:23,650 interesting. Part of the process was. 241 00:13:23,700 --> 00:13:27,090 Trying to figure out. How to clarify? 242 00:13:27,180 --> 00:13:31,630 What the difference is between plagiarism and the use of AI, R 243 00:13:31,700 --> 00:13:33,350 and I? Think a lot of that. 244 00:13:33,360 --> 00:13:37,390 Process really came down to a fundamental understanding. 245 00:13:37,400 --> 00:13:40,560 Of what AI? Does how it's different than 246 00:13:41,470 --> 00:13:47,360 like a plagiarizing or copying or anything else and trying to 247 00:13:47,370 --> 00:13:50,260 explain that? To people who had issues. 248 00:13:50,270 --> 00:13:53,180 You're amazing. I got the chance to explore it a 249 00:13:53,190 --> 00:13:57,580 little bit in the forward explores the traditional values 250 00:13:57,590 --> 00:14:00,940 of scholarship, so talks about authorship, reliability, efforts 251 00:14:00,950 --> 00:14:04,300 and merit and how AI interacts with each of those. 252 00:14:04,310 --> 00:14:08,740 And you also introduced in your in the publication A5 level 253 00:14:08,750 --> 00:14:12,500 taxonomy to classify I involvement in legal writing, 254 00:14:12,510 --> 00:14:16,450 which I thought was very, very interesting and quite a 255 00:14:16,460 --> 00:14:19,980 scholarly work on itself. So was there, you know, a few 256 00:14:19,990 --> 00:14:23,640 references that that led you to structure in that way? 257 00:14:23,710 --> 00:14:28,120 And now, how have you heard from authors, editors, professors? 258 00:14:28,240 --> 00:14:30,610 How have they responded to that system so far? 259 00:14:30,660 --> 00:14:36,050 Yeah, so the the creation of the five level taxonomy mirrors the 260 00:14:36,120 --> 00:14:39,990 Society for Automotive Engineering level of automation. 261 00:14:40,060 --> 00:14:43,030 Like I alluded to earlier, my background before law school was 262 00:14:43,040 --> 00:14:45,710 in science policy. And one thing that I worked a 263 00:14:45,720 --> 00:14:50,090 lot on was researchers who work was supporting researchers who 264 00:14:50,100 --> 00:14:52,650 were trying to get funding for automated vehicle projects. 265 00:14:52,720 --> 00:14:56,590 And so I was relatively familiar with the SAE levels because it's 266 00:14:56,600 --> 00:14:59,610 something that's been broadly incorporated into a lot of the 267 00:14:59,620 --> 00:15:02,350 governance of automated technology in the United States 268 00:15:02,360 --> 00:15:04,850 and probably around the world, but I only know about the United 269 00:15:04,860 --> 00:15:06,030 States. Someone going to talk about 270 00:15:06,040 --> 00:15:07,630 that. But that provided a really 271 00:15:07,640 --> 00:15:11,230 valuable anchoring point for trying to find a way to put 272 00:15:11,240 --> 00:15:14,130 language to the way that we're interacting with it. 273 00:15:14,140 --> 00:15:16,870 I thought that a human, what I would call a human factors 274 00:15:16,880 --> 00:15:20,710 approach to this kind of taxonomy was really valuable 275 00:15:20,780 --> 00:15:23,390 because it felt like, particularly in light of the 276 00:15:23,400 --> 00:15:26,460 paradigm that we kind of announced and then tried to, you 277 00:15:26,470 --> 00:15:30,340 know, analyze about authorship and merit and, and all of these 278 00:15:30,350 --> 00:15:31,970 things. Is that at the core of that, 279 00:15:31,980 --> 00:15:33,750 it's the work that the human is doing. 280 00:15:33,840 --> 00:15:39,390 And so the taxonomy tries to be a first bite at the apple to 281 00:15:39,440 --> 00:15:45,090 make the point when it comes to disclosing and, and telling 282 00:15:45,100 --> 00:15:48,530 someone about the role that I played is really anchoring it to 283 00:15:48,540 --> 00:15:51,270 what did the human do and how involved was the human? 284 00:15:51,320 --> 00:15:55,270 And one of our approaches there was that, you know, if we make 285 00:15:55,280 --> 00:15:59,200 it about the product that's used, that's going to be, you 286 00:15:59,210 --> 00:16:01,730 know, that's going to be, you know, out of service in a week 287 00:16:01,740 --> 00:16:04,430 because they're gonna have new usages, they're gonna have new, 288 00:16:04,480 --> 00:16:06,790 new use cases for AI in a second. 289 00:16:06,800 --> 00:16:09,190 And so there's nothing Evergreen about that. 290 00:16:09,200 --> 00:16:12,090 And so this was our approach to kind of provide some kind of 291 00:16:12,100 --> 00:16:16,540 Evergreen kind of taxonomy. And so that was, I think all of 292 00:16:16,550 --> 00:16:19,850 those were kind of the loadstar that kind of guided the 293 00:16:19,860 --> 00:16:22,610 formation of this. I think the, the feedback that 294 00:16:22,620 --> 00:16:25,370 we've gotten from scholars that we've, we've taken it to has 295 00:16:25,380 --> 00:16:29,930 been really positive. Our AI advisors really was on 296 00:16:29,940 --> 00:16:32,110 board with it. The, the authors that we worked 297 00:16:32,120 --> 00:16:35,670 with, professors Andrew Torrance and, and Bill Tomlinson, who are 298 00:16:35,680 --> 00:16:41,280 kind of pushing the, the pushing edge of this exploration, um, of 299 00:16:41,290 --> 00:16:44,550 AI legal scholarship. We're very excited about it. 300 00:16:44,560 --> 00:16:47,360 It, it kind of coincides with the work that they've done that 301 00:16:47,370 --> 00:16:51,360 they call like organic websites, trying to borrow from organic 302 00:16:51,370 --> 00:16:55,480 produce regulations to try and be able to taxonomies the way 303 00:16:55,490 --> 00:16:58,880 that automated content is, is delivered on websites. 304 00:16:58,890 --> 00:17:00,260 It's something that they've written about. 305 00:17:00,330 --> 00:17:04,540 And so this model of borrowing from other kinds of regulatory 306 00:17:04,550 --> 00:17:08,140 schemes or disclosure schemes is something that, that I think is 307 00:17:08,150 --> 00:17:12,720 a pretty well accepted practice as we enter into new and 308 00:17:12,730 --> 00:17:16,180 emerging technological fields. The one, the one piece of 309 00:17:16,190 --> 00:17:18,740 critique that we've seen, I think the above the law article 310 00:17:18,750 --> 00:17:21,579 kind of enunciates this critique pretty well is that, you know, 311 00:17:21,589 --> 00:17:24,599 on this sliding scale, it does become kind of difficult at 312 00:17:24,609 --> 00:17:28,910 times to know when you're jumping from level 2 to Level 3 313 00:17:28,920 --> 00:17:31,260 to level 4. The only real bright lines are 314 00:17:31,270 --> 00:17:35,020 level 1 and level 5 because level 1 is no AI at all. 315 00:17:35,070 --> 00:17:38,420 Level 5 is only AI all the time. And then there's this 316 00:17:38,430 --> 00:17:41,010 intermediate point. And I think that that critiques, 317 00:17:41,020 --> 00:17:43,580 right. But I also think that that's 318 00:17:43,590 --> 00:17:47,360 something that can be molded and perhaps tooled with. 319 00:17:47,370 --> 00:17:50,300 Perhaps having one intermediary level is the solution to that 320 00:17:50,310 --> 00:17:53,040 problem. I think that the bigger step in 321 00:17:53,050 --> 00:17:56,110 our, what I would say, the better contribution of us taking 322 00:17:56,120 --> 00:17:59,340 that effort in first place is that it's a stepping off point 323 00:17:59,350 --> 00:18:03,680 for being able to just frontline disclose the use of AI and 324 00:18:03,690 --> 00:18:07,260 describe the the human AI interaction in a way that's 325 00:18:07,270 --> 00:18:09,980 digestible the reader so that they know what they're reading. 326 00:18:09,990 --> 00:18:12,300 Because I think at the end of the day, like the most important 327 00:18:12,310 --> 00:18:14,880 thing here is that we're producing scholarship and that 328 00:18:14,890 --> 00:18:17,600 people know broadly what they're interacting with. 329 00:18:17,670 --> 00:18:22,340 Yeah, I think there were, there were two kind of policy concerns 330 00:18:22,390 --> 00:18:25,640 regarding the creation. Of the the five levels. 331 00:18:25,730 --> 00:18:29,380 And I think 1. Is the ethicality of publishing. 332 00:18:29,390 --> 00:18:32,440 AI at this stage in the. AI life cycle. 333 00:18:32,770 --> 00:18:36,880 I think it's important to just make. 334 00:18:36,890 --> 00:18:38,340 People aware, right? Because. 335 00:18:38,350 --> 00:18:40,860 Of some of the issues. That AI has. 336 00:18:40,930 --> 00:18:44,700 Especially when you're submitting into scholarship, I 337 00:18:44,710 --> 00:18:47,000 think the disclosure. Of the use of AI. 338 00:18:47,010 --> 00:18:49,400 Is important and then the other side. 339 00:18:49,590 --> 00:18:52,180 Is. Just kind of the human side of 340 00:18:52,270 --> 00:18:56,430 people have an easier time accepting what they understand. 341 00:18:56,440 --> 00:19:00,730 And so giving these categorizations, you know of how 342 00:19:00,740 --> 00:19:04,570 much AI is being used, what you can expect that means from a 343 00:19:04,580 --> 00:19:06,530 paper. About how it was drafted. 344 00:19:06,540 --> 00:19:10,010 How it was implemented, all those things, I think just gives 345 00:19:10,020 --> 00:19:15,930 people kind of an innate sense. Of comfortability as opposed to. 346 00:19:16,010 --> 00:19:18,290 Just kind of guessing. Is this AI? 347 00:19:18,300 --> 00:19:20,670 Is this not AI? I think just having that 348 00:19:20,680 --> 00:19:26,110 knowledge makes people more open to the concept in general. 349 00:19:26,200 --> 00:19:28,540 That's amazing. And there's so many different 350 00:19:28,550 --> 00:19:30,950 tools, so many possibilities of doing that. 351 00:19:30,960 --> 00:19:34,130 And I'm glad kind of it got broken into, but also I feel 352 00:19:34,140 --> 00:19:38,180 like the AI providers at the different companies that do, 353 00:19:38,190 --> 00:19:42,100 especially in the generative generative AI part of it, have 354 00:19:42,110 --> 00:19:46,930 also noticed this need. And, and I noticed that now this 355 00:19:46,940 --> 00:19:50,130 is not your recent, but yeah, within a couple of months of the 356 00:19:50,140 --> 00:19:53,440 development, for example, ChatGPT got the possibility of 357 00:19:53,450 --> 00:19:57,020 you like creating a specific link for that conversation you 358 00:19:57,030 --> 00:19:59,030 had so that you can share it in another. 359 00:19:59,040 --> 00:20:01,850 People can kind of go through the process and all that. 360 00:20:01,860 --> 00:20:05,900 So I feel like that does kind of align with the idea that we want 361 00:20:05,910 --> 00:20:09,210 to be transparent and then we we can share how we went about it. 362 00:20:09,220 --> 00:20:11,850 And I just really like that, that possibility. 363 00:20:12,660 --> 00:20:15,150 And then there's also a lot of different concerns. 364 00:20:15,200 --> 00:20:18,040 What we hear about AI in the legal field usually have to do 365 00:20:18,050 --> 00:20:21,390 with, you know, some somebody submitted a fact and or some 366 00:20:21,400 --> 00:20:24,870 kind of legal document into the court without checking, but what 367 00:20:24,880 --> 00:20:28,110 was the outcome of that document, which is which is so 368 00:20:28,120 --> 00:20:31,530 interesting because it it maybe. Could be compared to submitting 369 00:20:31,540 --> 00:20:34,470 something that was written by somebody who's not legally 370 00:20:34,480 --> 00:20:37,210 trained or, you know, a lost student or something like that. 371 00:20:37,220 --> 00:20:39,670 So it's always interesting to see that in that end. 372 00:20:39,680 --> 00:20:44,250 But there's this big concern about hallucinating the sources 373 00:20:44,260 --> 00:20:48,910 missing, misrepresenting the legal propositions or, you know, 374 00:20:49,040 --> 00:20:51,190 coming up with case law that never existed. 375 00:20:51,260 --> 00:20:54,730 So how do you train your editorial team to handle these? 376 00:20:54,800 --> 00:20:58,890 And also, what advice do you have to people who might be 377 00:20:58,940 --> 00:21:02,350 specifically concerned about the hallucination part of it? 378 00:21:02,400 --> 00:21:05,960 Yeah, so the. That's absolutely a concern and 379 00:21:05,970 --> 00:21:08,010 things like that do happen what I will. 380 00:21:08,020 --> 00:21:10,850 Say is that AI is. Getting better all the time and 381 00:21:10,860 --> 00:21:14,040 so it's becoming less and less frequent, although I 382 00:21:14,050 --> 00:21:18,570 hallucinations are absolutely still a problem in the AI 383 00:21:18,580 --> 00:21:21,130 community and the use of. AI as of today, but. 384 00:21:21,140 --> 00:21:24,130 There are much less than a few years ago. 385 00:21:24,200 --> 00:21:27,430 The way AI should be used is as a tool. 386 00:21:27,440 --> 00:21:29,710 It is not at the point and won't. 387 00:21:29,720 --> 00:21:34,190 Be for a while, where AI is the true substitute for human. 388 00:21:34,200 --> 00:21:39,170 Work thought, interaction. It needs to be targeted and it 389 00:21:39,180 --> 00:21:41,450 needs. To be checked as far as 390 00:21:41,600 --> 00:21:46,210 scholarship goes for. Publication the truth is that 391 00:21:46,260 --> 00:21:52,410 checking for existence, for correctness that they support 392 00:21:52,420 --> 00:21:55,320 the. Proposition that the words that. 393 00:21:55,380 --> 00:21:56,760 Have been submitted. To us. 394 00:21:56,770 --> 00:22:02,040 Are truthful and correct that's. Already part of the process of. 395 00:22:02,110 --> 00:22:05,260 A legal journal every day. Right, that's our job, whether 396 00:22:05,270 --> 00:22:07,680 it was written by. Humans, whether it was written 397 00:22:07,690 --> 00:22:12,540 by AI, no matter what, when we publish something as a journal, 398 00:22:12,590 --> 00:22:16,040 we are putting our name and things behind. 399 00:22:16,050 --> 00:22:18,240 It as far as the. Authenticity, which is the 400 00:22:18,250 --> 00:22:21,840 reason that legal journals have so much, so many hours. 401 00:22:21,850 --> 00:22:24,780 Of site. Checking and technical editing 402 00:22:24,790 --> 00:22:27,480 and all these things. Are for the purpose of. 403 00:22:27,490 --> 00:22:30,940 Minimizing and eliminating any of these errors that could come 404 00:22:30,950 --> 00:22:34,940 up and so as far as the difference between what it looks 405 00:22:34,950 --> 00:22:36,000 like. To edit. 406 00:22:36,410 --> 00:22:39,800 An article that was written. By or with the assistance. 407 00:22:39,870 --> 00:22:44,430 Of an eye and what it looks like to edit an article that was 408 00:22:44,440 --> 00:22:47,470 written entirely by a human the. Process. 409 00:22:47,480 --> 00:22:52,390 Itself is very very similar. I think the biggest point where 410 00:22:52,400 --> 00:22:54,650 that you see a divergences maybe the. 411 00:22:54,660 --> 00:23:02,160 Hours that it requires because. A human author might have a 412 00:23:02,170 --> 00:23:04,860 citation that doesn't support what they're saying, but maybe 413 00:23:04,870 --> 00:23:08,580 they meant for that citation to refer to something that was 414 00:23:08,590 --> 00:23:10,700 earlier in the paper or later in the paper. 415 00:23:10,750 --> 00:23:14,440 AI will completely hallucinate. They'll make up articles or 416 00:23:14,450 --> 00:23:17,260 sometimes cite to articles that don't necessarily support what 417 00:23:17,270 --> 00:23:20,200 they're saying because AI is trying to give you the answers, 418 00:23:20,210 --> 00:23:21,760 right? And so that's its goal. 419 00:23:21,770 --> 00:23:27,140 And so that is different between AI written and human written. 420 00:23:27,150 --> 00:23:31,680 And so there's a little bit more work that goes on to identify 421 00:23:31,690 --> 00:23:35,360 where AI is coming up with some of its information, where for a 422 00:23:35,370 --> 00:23:37,920 human written article, it's often more of a reorganization 423 00:23:37,930 --> 00:23:40,390 challenge as opposed. To OK, where did. 424 00:23:40,400 --> 00:23:43,480 This come from and there's also the and a challenge of if there 425 00:23:43,490 --> 00:23:45,840 is an issue with the human written thing, you can always 426 00:23:45,850 --> 00:23:48,280 reach out to the author and say hey sentence. 427 00:23:48,510 --> 00:23:51,670 Two of page, whatever. I don't think this is the site 428 00:23:51,680 --> 00:23:54,160 you meant to use. Is there something that you use 429 00:23:54,170 --> 00:23:56,820 from this and AI? It's much more. 430 00:23:56,830 --> 00:23:59,860 Difficult to try to get. That information to highlight 431 00:23:59,870 --> 00:24:02,280 one sentence, say, hey, where's this information coming from? 432 00:24:02,330 --> 00:24:05,080 Often I might just double down on the original. 433 00:24:05,090 --> 00:24:08,460 Citation it gave you. Yeah, I, I think that the 434 00:24:08,470 --> 00:24:12,740 hallucination issue, one of the articles that we rely on a lot 435 00:24:12,750 --> 00:24:18,080 in our piece is this Stanford study about the way in which AI 436 00:24:18,150 --> 00:24:20,820 operates when it's doing this kind of research. 437 00:24:20,830 --> 00:24:23,030 And one thing that they point out is that there's two kinds of 438 00:24:23,040 --> 00:24:25,580 hallucinations. There's the one that everyone 439 00:24:25,590 --> 00:24:28,900 talks about and gets you sanctioned in court, which is 440 00:24:28,910 --> 00:24:32,480 where the AI is made something up entirely out of whole cloth 441 00:24:32,490 --> 00:24:35,900 and it doesn't exist. The other one is where AI is, is 442 00:24:35,910 --> 00:24:39,080 stretching A proposition. And I think one thing that we 443 00:24:39,270 --> 00:24:45,980 tried to do was first creating a quicker way for staffers to be 444 00:24:45,990 --> 00:24:49,230 able to process those, which was creating a highlighting key 445 00:24:49,240 --> 00:24:54,600 basically for them to be able to streamline the identification of 446 00:24:54,610 --> 00:24:57,840 those items. Obviously our process still 447 00:24:57,850 --> 00:25:00,400 includes, you know, if you see something that doesn't support, 448 00:25:00,410 --> 00:25:03,670 then you go find something else that supports that proposition 449 00:25:03,730 --> 00:25:06,580 so that you can, you know, so that we can do our job as a 450 00:25:06,590 --> 00:25:09,350 journal and and reliably important things we're doing. 451 00:25:10,010 --> 00:25:13,430 But because of, and I think it's something that Spencer just 452 00:25:13,440 --> 00:25:18,640 alluded to is that the error attribution is different because 453 00:25:18,650 --> 00:25:22,560 when a human makes a mistake, we chalk it up to good faith. 454 00:25:22,570 --> 00:25:24,420 They tried, they just made a mistake. 455 00:25:24,470 --> 00:25:27,460 Whereas when you're dealing with new technology doing it, the 456 00:25:27,470 --> 00:25:32,000 errors entirely on the computer. And so because there's no like 457 00:25:32,010 --> 00:25:35,360 Direct Line of accountability to the person who made the the 458 00:25:35,370 --> 00:25:39,680 choice, they're become their opens up a lot more questions as 459 00:25:39,690 --> 00:25:42,380 to the reliability. But it also opens up a lot more 460 00:25:42,390 --> 00:25:47,400 solutions, which is that we can retailer the proposition of that 461 00:25:47,410 --> 00:25:50,770 sentence to support what the article does say, because we 462 00:25:50,780 --> 00:25:53,670 know that the article is right or we know the article exists. 463 00:25:53,760 --> 00:25:56,390 We know the article has said something that can support that 464 00:25:56,400 --> 00:25:59,670 that proposition authentically. And then the question also 465 00:25:59,680 --> 00:26:02,790 becomes like when you are looking at a paragraph full of 466 00:26:02,800 --> 00:26:06,660 propositions, when you, you know, knock one of those runs 467 00:26:06,670 --> 00:26:09,390 out of the ladder or you adjust it, does it still support the 468 00:26:09,400 --> 00:26:12,950 broader argument? And as long as the authors 469 00:26:12,960 --> 00:26:16,510 broader argument remains intact with like that little bit of 470 00:26:16,520 --> 00:26:19,030 tinkering so that it is supported. 471 00:26:19,100 --> 00:26:22,590 That's something that working with an AI with with a with 472 00:26:22,600 --> 00:26:25,470 scholarship that's created in conjunction with AI. 473 00:26:25,560 --> 00:26:29,470 Authors are more malleable in terms of the way that their pros 474 00:26:29,480 --> 00:26:33,270 is edited because in a lot of respects, they're less defensive 475 00:26:33,320 --> 00:26:37,520 of individual wording decisions. And so you do have a lot, you 476 00:26:37,530 --> 00:26:40,720 have a lot more problems because of the hallucination issue, but 477 00:26:40,730 --> 00:26:42,910 you do also have a lot more solutions. 478 00:26:42,960 --> 00:26:45,930 And so that was kind of an iterative process as we 479 00:26:45,980 --> 00:26:49,230 undertook this was identifying those different tools in our 480 00:26:49,240 --> 00:26:52,110 toolbox to be able to to do that. 481 00:26:52,120 --> 00:26:56,310 So that I think on the staff training question, it was really 482 00:26:56,320 --> 00:26:59,630 an ongoing process of identifying those where we can 483 00:26:59,720 --> 00:27:03,130 input more efficiencies in terms of like identification of issues 484 00:27:03,140 --> 00:27:05,420 that we can communicate to the author and put it in the authors 485 00:27:05,430 --> 00:27:07,510 court to deal with. Or where we can take more 486 00:27:07,520 --> 00:27:10,310 liberty knowing that the author is going to be more receptive to 487 00:27:10,320 --> 00:27:13,370 this kind of change. Because, you know, we're not, 488 00:27:13,440 --> 00:27:16,590 we're not changing their individualized decision to write 489 00:27:16,600 --> 00:27:19,210 the sentence this way. We are, we are working 490 00:27:19,220 --> 00:27:21,740 collaboratively with them and the product they have 491 00:27:21,790 --> 00:27:24,900 manufactured using these tools to make it the best possible 492 00:27:24,910 --> 00:27:26,900 product. And so I think that that side of 493 00:27:26,910 --> 00:27:29,400 the collaborative relationship is actually a little enhanced 494 00:27:29,440 --> 00:27:32,620 because the personal defensiveness of authors is is 495 00:27:32,630 --> 00:27:35,440 actually muted a little bit. And I think it makes sense. 496 00:27:35,450 --> 00:27:38,940 Like authors are for the. Traditional legal scholarship 497 00:27:38,950 --> 00:27:43,560 authors are spending hundreds of hours doing research, framing 498 00:27:43,570 --> 00:27:45,760 every sentence and so. These. 499 00:27:45,830 --> 00:27:49,330 Papers and articles become passion projects and just like 500 00:27:49,340 --> 00:27:51,740 any passion project, you you become. 501 00:27:51,750 --> 00:27:55,810 Very attached to the choices. You made to get yourself along 502 00:27:55,820 --> 00:27:58,500 the way what Coop is referring. To is like the stylistic. 503 00:27:58,510 --> 00:28:03,200 Kind of changes that don't affect the maybe the meaning of 504 00:28:03,210 --> 00:28:07,740 what's going on, but might enhance readability when you've. 505 00:28:07,750 --> 00:28:12,090 Used the tools as coop said, like AI to kind of help make 506 00:28:12,100 --> 00:28:15,880 this project first you probably. Haven't spend quite as many 507 00:28:15,890 --> 00:28:17,790 hours because you weren't physically typing out the words. 508 00:28:17,800 --> 00:28:20,430 That's not to say you didn't do all the research and all that 509 00:28:20,440 --> 00:28:24,290 kind of stuff, but AI, one of the ways that it helps is it can 510 00:28:24,360 --> 00:28:27,890 give you a first draft. Or it can give you. 511 00:28:27,940 --> 00:28:31,630 Something where the physical typing of every word is not 512 00:28:31,640 --> 00:28:34,630 something you have to do and. So because of that. 513 00:28:34,920 --> 00:28:39,190 If a journal suggests a change to something stylistic, that was 514 00:28:39,200 --> 00:28:41,070 done. By the AI. 515 00:28:41,140 --> 00:28:44,330 Right, as opposed to an author, there's there's less attachment. 516 00:28:44,340 --> 00:28:47,030 To it just because there. Isn't the same level of. 517 00:28:47,320 --> 00:28:50,670 Commitment to exactly the way that that was written. 518 00:28:50,680 --> 00:28:53,670 And so it gives journals more freedom. 519 00:28:53,720 --> 00:28:56,280 As far as shaping paper stylistically. 520 00:28:56,290 --> 00:28:59,550 Which is not something that's common for legal scholarship 521 00:28:59,560 --> 00:29:01,930 generally. Usually the authors are the 522 00:29:01,940 --> 00:29:04,320 driving force. Behind the. 523 00:29:04,330 --> 00:29:06,450 Style of a paper. Perfect. 524 00:29:06,460 --> 00:29:08,210 No, I love, I love that you mentioned that. 525 00:29:08,300 --> 00:29:11,110 And I'm thinking about a few different topics here too on 526 00:29:11,120 --> 00:29:14,010 the, on the sense of accessibility to league of 527 00:29:14,020 --> 00:29:19,130 scholarship and then even like for for foreign actors to get 528 00:29:19,140 --> 00:29:22,490 into North American legal scholarship, right. 529 00:29:22,580 --> 00:29:26,030 So there is a very specific way that legal scholarship is 530 00:29:26,040 --> 00:29:29,490 written and is portrayed. And sometimes, as you know, 531 00:29:29,550 --> 00:29:32,790 speakers of English as a second language or, you know, all kinds 532 00:29:32,800 --> 00:29:35,280 of different barriers that you might have for legal 533 00:29:35,290 --> 00:29:37,440 scholarship. This could be tools really that 534 00:29:37,510 --> 00:29:39,800 that level, that playing field in many ways. 535 00:29:39,870 --> 00:29:44,420 So in your view, does AI lower that barrier to enter in legal 536 00:29:44,430 --> 00:29:47,650 scholarship? But also does it possibly create 537 00:29:47,660 --> 00:29:51,740 some risk of, of having a divide between people who use the 538 00:29:51,750 --> 00:29:54,280 tools, people who don't, and then, you know, you have some 539 00:29:54,290 --> 00:29:57,680 kind of discredit or this merit for being in one of those 540 00:29:57,690 --> 00:29:59,480 groups? What are some of the the 541 00:29:59,490 --> 00:30:01,080 thoughts and things you've heard? 542 00:30:01,170 --> 00:30:06,930 Yeah, When I was in DC, I worked a lot on not only automation, 543 00:30:06,940 --> 00:30:08,850 automated vehicles, but biotechnology. 544 00:30:08,860 --> 00:30:12,270 So there's a point here. When it comes to biotechnology 545 00:30:12,280 --> 00:30:14,840 in the United States, it's historically been governed by 546 00:30:14,850 --> 00:30:18,110 something called the Coordinated Framework for Biotechnology, 547 00:30:18,180 --> 00:30:22,050 which in a roundabout way, its entire function is ensuring that 548 00:30:22,060 --> 00:30:24,650 you regulate products, not processes. 549 00:30:24,720 --> 00:30:29,580 And I think that the best outcome of this duality between 550 00:30:29,590 --> 00:30:32,610 those who use these tools and those who don't is that we get 551 00:30:32,620 --> 00:30:35,770 to a point where law reviews are evaluating products on their 552 00:30:35,780 --> 00:30:38,530 merit and not the way in which they were created. 553 00:30:38,580 --> 00:30:41,450 Because I do think that the point that you raised there is 554 00:30:41,460 --> 00:30:45,480 that this does level a lot of the playing field because not 555 00:30:45,490 --> 00:30:49,250 only does it help folks who maybe don't speak English or are 556 00:30:49,260 --> 00:30:53,030 unfamiliar with the structure of a law review to be able to, you 557 00:30:53,040 --> 00:30:56,990 know, run it through Claude or something to either translate it 558 00:30:57,040 --> 00:31:03,010 into a workable translation or reformat it into a more law 559 00:31:03,020 --> 00:31:07,990 review type esque headers and, and structure and everything. 560 00:31:08,000 --> 00:31:10,950 But I think it also levels the playing field on infrastructure 561 00:31:11,000 --> 00:31:15,420 because right now you have a huge duality between individual 562 00:31:15,430 --> 00:31:19,130 researchers and authors and researchers, research professors 563 00:31:19,140 --> 00:31:24,190 who are well funded, who have, you know, access to work from, 564 00:31:24,200 --> 00:31:27,530 from RA and other support staff. And in a lot of ways, I think 565 00:31:27,540 --> 00:31:30,390 this kind of hearkens back to a conversation we had about the 566 00:31:30,400 --> 00:31:33,810 book, the legal Singularity. And the way that AI allows for 567 00:31:33,820 --> 00:31:37,050 the democratization of legal information is that it also 568 00:31:37,060 --> 00:31:40,430 allows for the democratization of intellectual labour. 569 00:31:40,500 --> 00:31:45,190 And so where researchers or authors are able to offload a 570 00:31:45,200 --> 00:31:48,770 lot of the cognitive burden or just the research work using 571 00:31:48,780 --> 00:31:52,330 tools like illicit, which will do like a full, you know, 572 00:31:52,340 --> 00:31:55,930 research project for you on research outline or using it to 573 00:31:56,020 --> 00:31:59,730 perhaps outline the argument structure of various case law 574 00:31:59,740 --> 00:32:02,170 and things like that. They're able to offload the work 575 00:32:02,180 --> 00:32:05,350 that someone would just be handing off to another person 576 00:32:05,420 --> 00:32:08,050 anyway. And so you're able to level the 577 00:32:08,060 --> 00:32:09,750 playing field in a lot of respects there. 578 00:32:09,800 --> 00:32:12,590 I, I think for the time being, there's going to continue to be 579 00:32:12,600 --> 00:32:15,720 some form of discriminatory pattern with respect to, you 580 00:32:15,730 --> 00:32:19,000 know, there's gonna always be people skeptical of the adoption 581 00:32:19,010 --> 00:32:21,100 of technology or the usage of technology. 582 00:32:21,110 --> 00:32:23,780 I mean, for a long time, you know, people weren't allowed to 583 00:32:23,790 --> 00:32:27,940 wear pants with zippers because it was provocative to the old 584 00:32:27,950 --> 00:32:29,640 guard. In the same way lawyers weren't 585 00:32:29,650 --> 00:32:34,040 allowed to use the Internet or word processor or not cite 586 00:32:34,050 --> 00:32:36,190 something that they couldn't point to the exact book that 587 00:32:36,200 --> 00:32:39,260 they pulled it from. And so I think that tradition of 588 00:32:39,270 --> 00:32:43,060 skepticism is going to carry forward in academia and law. 589 00:32:43,110 --> 00:32:46,480 But I do think that this does level the playing field in a lot 590 00:32:46,490 --> 00:32:49,600 of really meaningful ways. And as long as the paradigm 591 00:32:49,610 --> 00:32:53,380 shifts towards product rather than process, I think that 592 00:32:53,390 --> 00:32:56,260 overcomes a lot of that innate skepticism. 593 00:32:56,410 --> 00:33:00,040 Yeah, I, I agree completely. I think that the skepticism is 594 00:33:00,110 --> 00:33:04,300 is here and it's going to. Be here for probably a while. 595 00:33:04,310 --> 00:33:07,200 It's it's something you see with the adoption of every type of 596 00:33:07,210 --> 00:33:12,480 new technology that's ever been invented, there is, you know, a 597 00:33:12,490 --> 00:33:15,760 kind of a period of turmoil while, you know. 598 00:33:16,010 --> 00:33:20,660 The world rigor tests. The new technology, right? 599 00:33:20,670 --> 00:33:23,180 It happened, Scoop said. With the Internet. 600 00:33:23,190 --> 00:33:25,140 Where people were very skeptical. 601 00:33:25,150 --> 00:33:27,940 Of, you know, cases that that were coming from the Internet 602 00:33:27,950 --> 00:33:31,970 because for so long it was from a book or it didn't exist. 603 00:33:31,980 --> 00:33:35,420 And so there was kind of that skepticism of can I trust this? 604 00:33:35,430 --> 00:33:37,380 OK, I found it on the Internet. Now let me go find it in the 605 00:33:37,390 --> 00:33:40,560 book just so I can make. Sure that it's it's that there 606 00:33:40,570 --> 00:33:44,080 is, so there's. That level of skepticism, I 607 00:33:44,090 --> 00:33:49,160 think it's healthy to a certain extent, especially while we're 608 00:33:49,170 --> 00:33:51,260 in this stage. Of AI where there are. 609 00:33:51,270 --> 00:33:54,660 Hallucinations and there are reasons for, you know, well 610 00:33:54,670 --> 00:33:58,040 founded reasons for skepticism. So I think that part is healthy. 611 00:33:58,090 --> 00:34:02,140 I think it will diminish more and more and more, and I think 612 00:34:02,190 --> 00:34:07,580 that it may not be universal adoption in the near. 613 00:34:07,590 --> 00:34:11,100 Future It will probably take a while before there's. 614 00:34:11,170 --> 00:34:13,940 You know, an overwhelming. Majority who are. 615 00:34:13,989 --> 00:34:17,500 At least vocal or open. About the use of AI. 616 00:34:17,510 --> 00:34:19,219 And legal work and legal scholarship. 617 00:34:19,230 --> 00:34:22,300 And other fields just because. In many cases, like the 618 00:34:22,310 --> 00:34:24,560 familiarity of using things the way we. 619 00:34:24,570 --> 00:34:28,440 Have is. So it's so much familiarity. 620 00:34:28,449 --> 00:34:30,570 With that that it's hard to give it. 621 00:34:30,580 --> 00:34:34,710 Up the skepticism, and also there is a bit of a learning 622 00:34:34,719 --> 00:34:38,850 curve that comes to using AI, especially when you're trying. 623 00:34:39,139 --> 00:34:43,110 To purposefully use it. Responsibly and ethically and 624 00:34:43,120 --> 00:34:44,550 so. There is a bit of that. 625 00:34:44,600 --> 00:34:47,210 Learning curve that not everybody is going to want to 626 00:34:47,219 --> 00:34:51,330 attack and so the nice thing. Is just like with the. 627 00:34:51,380 --> 00:34:55,650 Advents of other technologies right the people who don't want 628 00:34:55,659 --> 00:34:59,350 to use AI in their scholarship or skeptical of it or think it's 629 00:34:59,360 --> 00:35:03,560 really important to do everything where they're typing 630 00:35:03,570 --> 00:35:05,850 the words where they're. Using RA's for. 631 00:35:05,900 --> 00:35:08,670 Some of the research and things, hey, it doesn't propose any 632 00:35:08,680 --> 00:35:11,690 like. Blockade of that old style. 633 00:35:11,740 --> 00:35:14,500 Or the current style, the present way of doing things. 634 00:35:14,510 --> 00:35:16,910 There's no it doesn't make any of that any. 635 00:35:16,920 --> 00:35:20,060 Harder, it just gives a new path for. 636 00:35:20,200 --> 00:35:24,410 People who don't have perhaps the time, the financial. 637 00:35:24,420 --> 00:35:29,690 Resources or elsewise to. To take on that labour, it 638 00:35:29,740 --> 00:35:34,250 levels the playing field so that they can engage in this 639 00:35:34,260 --> 00:35:36,110 intellectual conversation as well. 640 00:35:36,180 --> 00:35:38,520 That's fantastic. Well, thank you so much for 641 00:35:38,600 --> 00:35:42,430 sharing those insights and and I really appreciate the way that 642 00:35:42,440 --> 00:35:46,130 this can be a tool really to make so many things better. 643 00:35:46,140 --> 00:35:48,590 I like that that the proposition, since you mentioned 644 00:35:48,600 --> 00:35:52,940 the legal singularity, basically has the idea of making law 645 00:35:52,950 --> 00:35:56,040 significantly better. And I think that that just one 646 00:35:56,050 --> 00:35:59,520 of the ways is, is this very simple way of making it possible 647 00:35:59,530 --> 00:36:01,970 for other voices to be brought up and all of that. 648 00:36:02,070 --> 00:36:05,800 And finally, to wrap up today's episode, If I Lost you and our 649 00:36:05,810 --> 00:36:09,560 young scholar, once you write with the help of AI responsibly 650 00:36:09,570 --> 00:36:12,990 and ethically, I wanted to know what are some, you know, maybe 651 00:36:13,000 --> 00:36:16,680 top three, top five golden rules that may help them to get 652 00:36:16,690 --> 00:36:18,020 started. I think. 653 00:36:18,110 --> 00:36:21,780 I'll give the classic law a law. School answer of it depends if 654 00:36:21,790 --> 00:36:23,960 they're trying to publish something, I think the first 655 00:36:23,970 --> 00:36:28,540 step has to be you know, check what your drafting for right If 656 00:36:28,550 --> 00:36:30,060 you're. Drafting for. 657 00:36:31,670 --> 00:36:34,820 A publication that has an explicit ban on all AI assisted 658 00:36:34,830 --> 00:36:37,700 materials or something like. That it's going to be difficult 659 00:36:37,710 --> 00:36:39,820 to use. AI ethically in that context 660 00:36:39,830 --> 00:36:43,540 because there's an outright ban. And so, you know, in that 661 00:36:43,550 --> 00:36:45,600 situation you might want to stay away from it. 662 00:36:45,670 --> 00:36:51,100 But I think one of the most important things is just don't 663 00:36:51,110 --> 00:36:53,860 rely on AI. As the AI shouldn't be the 664 00:36:53,870 --> 00:36:58,120 backstop right? AI is a great tool but just like 665 00:36:58,130 --> 00:37:00,200 a Google search right? If you. 666 00:37:00,210 --> 00:37:03,900 Click on the very first link and do no further checking. 667 00:37:03,950 --> 00:37:06,760 You're not necessarily going to get the best information out of 668 00:37:06,770 --> 00:37:08,940 it, right? You don't have really any 669 00:37:08,950 --> 00:37:12,850 safeguards on whether that's correct, on whether there's 670 00:37:12,860 --> 00:37:14,650 something wrong. With the way that. 671 00:37:14,730 --> 00:37:16,420 They came to their conclusions or anything else. 672 00:37:16,430 --> 00:37:18,530 Like that? AI is the exact same way. 673 00:37:18,540 --> 00:37:22,790 If you type in one prompt and just take whatever response AI 674 00:37:22,800 --> 00:37:25,750 gives you and run with it, you leave yourself open to 675 00:37:25,760 --> 00:37:28,160 experiencing problems like the hallucinations that. 676 00:37:28,170 --> 00:37:30,030 We talked. About like some of the stretched 677 00:37:30,040 --> 00:37:34,350 propositions, and so anybody who's using it has a 678 00:37:34,360 --> 00:37:38,150 responsibility to check for correctness. 679 00:37:38,160 --> 00:37:41,650 Right. AI should not be used in a way 680 00:37:41,660 --> 00:37:46,310 that's just relies entirely on this new technology to come up 681 00:37:46,320 --> 00:37:50,070 with every aspect of an argument with no fact checking whatever. 682 00:37:50,140 --> 00:37:53,590 And that's probably one of the biggest problems that we see 683 00:37:53,600 --> 00:37:56,170 from things that are drafted by AI. 684 00:37:57,470 --> 00:38:00,050 I think building on that, my first one would be like 685 00:38:00,100 --> 00:38:02,710 accountability. You know, at the end of the day, 686 00:38:02,760 --> 00:38:04,910 the authors name is still on this work. 687 00:38:04,920 --> 00:38:08,130 We have, at least in in US intellectual property, 688 00:38:08,220 --> 00:38:11,650 consistently rejected the notion that that AI has an ownership or 689 00:38:11,660 --> 00:38:15,150 authorship right? And so as long as it's your name 690 00:38:15,160 --> 00:38:17,330 on the work, you have a responsibility to ensure that 691 00:38:17,340 --> 00:38:21,590 it's accurate and meritorious. I also think that AI should not 692 00:38:21,600 --> 00:38:24,050 create a shortcut for people to publish for the sake of 693 00:38:24,060 --> 00:38:26,190 publication. I think that scholarship is an 694 00:38:26,200 --> 00:38:30,250 ecosystem that deserves discretion in the way that 695 00:38:30,260 --> 00:38:33,750 people enter into it and attempt to participate in it. 696 00:38:33,840 --> 00:38:38,330 And so just because AI makes it possible for you to write an 697 00:38:38,340 --> 00:38:41,300 article doesn't mean you should be writing an article. 698 00:38:41,430 --> 00:38:43,580 Like at the end of the day, like that article still needs to mean 699 00:38:43,590 --> 00:38:45,680 something for someone. And I think particularly in 700 00:38:45,690 --> 00:38:50,000 legal scholarship, where our entire advance here is to, you 701 00:38:50,010 --> 00:38:53,640 know, in a in a path that that's somewhat parallels litigation 702 00:38:53,650 --> 00:38:56,600 and the common law, be able to explore valuable ideas. 703 00:38:56,690 --> 00:39:00,660 And a lot of times, and a lot of times not those ideas end up in 704 00:39:00,670 --> 00:39:04,700 briefings and in opinions and really materially affect the 705 00:39:04,710 --> 00:39:07,680 lives of litigants and people trying to vindicate themselves 706 00:39:07,690 --> 00:39:10,040 legally. It's important that, you know, 707 00:39:10,050 --> 00:39:12,720 you do a gut check on and make sure that what you're writing 708 00:39:12,730 --> 00:39:15,670 actually matters. And then you allow AI to come 709 00:39:15,680 --> 00:39:19,240 alongside you to move the ball down the field in pursuit of 710 00:39:19,250 --> 00:39:22,320 that objective. So I think like AI creates a lot 711 00:39:22,330 --> 00:39:26,350 of liberty and so far as it helps people do what would 712 00:39:26,360 --> 00:39:28,810 otherwise be incredibly cumbersome, but it shouldn't be 713 00:39:28,820 --> 00:39:31,370 a license to participate for the sake of participating. 714 00:39:31,380 --> 00:39:33,290 Fantastic. Well, I really appreciate your 715 00:39:33,300 --> 00:39:36,010 insights and really appreciate the work you put out there. 716 00:39:36,060 --> 00:39:39,130 And we'll definitely have the links down here so people can 717 00:39:39,140 --> 00:39:41,390 take a look. We really appreciate you both 718 00:39:41,400 --> 00:39:43,470 coming into the podcast today. Thank you so much. 719 00:39:44,260 --> 00:39:46,050 Thanks man. Thanks for having us.