1 00:00:00,280 --> 00:00:03,320 Speaker 1: Let's go first to build for asset management. Debra Lambie 2 00:00:03,360 --> 00:00:06,760 Speaker 1: of course, high Debra, Hi, how are you well? Thank you? 3 00:00:06,880 --> 00:00:09,799 Speaker 1: So how is AI being used to discover new medicines? 4 00:00:10,600 --> 00:00:13,039 Speaker 1: So it's interesting. I'm just recently back from a US 5 00:00:13,119 --> 00:00:16,440 Speaker 1: trip where I visited over forty different global healthcare companies, 6 00:00:16,680 --> 00:00:18,520 Speaker 1: and one of the things that really stuck out to 7 00:00:18,560 --> 00:00:21,600 Speaker 1: me was actually just how excited companies and investors are 8 00:00:21,880 --> 00:00:25,640 Speaker 1: about the potential of AI to transform the drug discovery process. 9 00:00:25,920 --> 00:00:29,000 Speaker 1: And that's the process that identifies new molecules that have 10 00:00:29,120 --> 00:00:32,400 Speaker 1: the potential to be used as a medicine. So artificial 11 00:00:32,400 --> 00:00:35,159 Speaker 1: intelligence models that you or I might use, these are 12 00:00:35,200 --> 00:00:38,400 Speaker 1: trained on language and generate a language output, and a 13 00:00:38,440 --> 00:00:41,480 Speaker 1: good example of this would be chat GPT, But then 14 00:00:41,600 --> 00:00:44,920 Speaker 1: there's also drug GPT, and this is an example of 15 00:00:44,960 --> 00:00:48,440 Speaker 1: an AI model which has trained on huge data sets 16 00:00:48,440 --> 00:00:51,440 Speaker 1: of known molecules and its output is to help discover 17 00:00:51,640 --> 00:00:54,480 Speaker 1: new molecules. And then this is really exciting because it 18 00:00:54,480 --> 00:00:57,240 Speaker 1: has the potential to make the drug discovery process much 19 00:00:57,360 --> 00:01:00,560 Speaker 1: faster and cheaper. Yeah, what is it that the traditional 20 00:01:00,600 --> 00:01:04,800 Speaker 1: drug discovery process is so time consuming and expensive? Exactly, 21 00:01:04,959 --> 00:01:08,160 Speaker 1: So for a medicine to be approved for use in patients. 22 00:01:08,200 --> 00:01:10,240 Speaker 1: All medicines have to go to the time triming and 23 00:01:10,319 --> 00:01:13,960 Speaker 1: expensive process of completing phase one, phase two, and place 24 00:01:14,040 --> 00:01:17,200 Speaker 1: three clinical trials to demonstrate that a medicine is both 25 00:01:17,280 --> 00:01:20,679 Speaker 1: safe and importantly effective in people. And this cost on 26 00:01:20,720 --> 00:01:23,920 Speaker 1: average over two billion dollars per new molecule and takes 27 00:01:23,959 --> 00:01:26,880 Speaker 1: on average around ten years. And then the success rate 28 00:01:26,920 --> 00:01:29,320 Speaker 1: isn't great either, with many medicines not making it all 29 00:01:29,360 --> 00:01:32,120 Speaker 1: the way through the process and in fact, on average 30 00:01:32,160 --> 00:01:34,840 Speaker 1: only around seven percent of molecules make it all the 31 00:01:34,840 --> 00:01:38,080 Speaker 1: way through from phase one clinical trials to approval. And 32 00:01:38,160 --> 00:01:41,560 Speaker 1: so if artificial intelligence can help bring the cost of 33 00:01:41,640 --> 00:01:44,440 Speaker 1: drug discovery down, that can only be a good things 34 00:01:44,480 --> 00:01:47,080 Speaker 1: that make governments and healthcare systems all around the world, 35 00:01:47,280 --> 00:01:50,040 Speaker 1: including here in New Zealans, are dealing with increasing costs 36 00:01:50,040 --> 00:01:51,960 Speaker 1: of healthcare. So right, yeah, how is it actually doing it? 37 00:01:52,000 --> 00:01:54,200 Speaker 1: How does AI bring down the cost and the time 38 00:01:54,480 --> 00:01:58,440 Speaker 1: of developing these new medicines. So it's really complicated to 39 00:01:58,520 --> 00:02:01,800 Speaker 1: design new molecules, and what we're seeing is that artificial 40 00:02:01,840 --> 00:02:05,640 Speaker 1: intelligence can go to the design process of new molecules 41 00:02:05,680 --> 00:02:09,560 Speaker 1: approximately ten times faster, and then it can run simulations 42 00:02:09,639 --> 00:02:12,959 Speaker 1: or testing on those molecules about one hundred times faster 43 00:02:13,040 --> 00:02:16,360 Speaker 1: compared to traditional labs, and this reduces the time and 44 00:02:16,400 --> 00:02:19,799 Speaker 1: the cost involved in developing new medicines. And so if 45 00:02:19,800 --> 00:02:21,880 Speaker 1: we think about how companies are doing this, you will 46 00:02:22,000 --> 00:02:24,920 Speaker 1: heard of the concept of self driving cars. Then in 47 00:02:24,960 --> 00:02:27,680 Speaker 1: the world of farmer there's also the concept of self 48 00:02:27,720 --> 00:02:31,520 Speaker 1: driving labs. So these labs are using artificial intelligence and 49 00:02:31,720 --> 00:02:36,680 Speaker 1: robotics combined to run a cycle of prediction of new molecules, 50 00:02:36,760 --> 00:02:40,240 Speaker 1: experimentation on those new molecules, and then analysis of those 51 00:02:40,240 --> 00:02:44,600 Speaker 1: molecules to identify most promising compounds that are then tested 52 00:02:44,720 --> 00:02:46,640 Speaker 1: in the real world. Who are the companies the main 53 00:02:46,680 --> 00:02:50,000 Speaker 1: companies who are involved here, So a good example another 54 00:02:50,080 --> 00:02:52,600 Speaker 1: company that helps with self driving labs so US company 55 00:02:52,600 --> 00:02:56,480 Speaker 1: called Betton Dickinson and it's leader in lab automation systems 56 00:02:56,800 --> 00:02:59,400 Speaker 1: and its technology can take out around a third of 57 00:02:59,440 --> 00:03:01,919 Speaker 1: the cost of the traditional lab. And then on the 58 00:03:02,000 --> 00:03:04,640 Speaker 1: drug discovery side, what we're seeing is a number of 59 00:03:04,680 --> 00:03:08,480 Speaker 1: AI focused companies of partnering with big pharmaceutical companies. So 60 00:03:08,800 --> 00:03:13,480 Speaker 1: for example, there's a UKAI company called Benevolent Ai and 61 00:03:13,520 --> 00:03:16,640 Speaker 1: it's partners with Astra Zeneca, which is a large farmer company, 62 00:03:16,840 --> 00:03:19,280 Speaker 1: and Astra Zenker is using its platform to find new 63 00:03:19,280 --> 00:03:23,000 Speaker 1: ways to treat disease and personalized medicine insipation. Debraah, the 64 00:03:23,040 --> 00:03:25,160 Speaker 1: stuff is absolutely fascinating, isn't it. Thank you so much 65 00:03:25,200 --> 00:03:28,080 Speaker 1: for talking us through this. Deborah Landy Milfed Asset Management. 66 00:03:28,800 --> 00:03:31,959 Speaker 1: For more from Heather Duplasy Allen Drive, listen live to 67 00:03:32,080 --> 00:03:35,200 Speaker 1: News Talks B from four pm weekdays, or follow the 68 00:03:35,280 --> 00:03:36,920 Speaker 1: podcast on iHeartRadio.