On June 21, 2018, LaunchBio hosted an event focused on artificial intelligence (AI) at Genesis in Campus Point. Panelists were Dr. Khai Minh Pham, founder of ThinkingNode; Thomas Kluz, Head of Healthcare Investing at Qualcomm Ventures and General Partner of dRx Capital; Navid Alipour, Co-founder and Managing Partner of Analytics Ventures; Casey Laris, CTO of Reveal Biosciences. The session was moderated by Margot Wohl, a graduate student in neuroscience at UCSD and KPBS Podcast producer.

After the opening reception, Rebecca Beattie welcomed the attendees and Dr. Khai Minh Pham gave the opening remarks. AI is being incorporated in drug discovery more and more. China, in particular, invests heavily in AI. Start-ups, investors, and companies with AI expertise have different skills, interests, and goals but, for AI to be an integral part of drug discovery, these 3 elements need to be combined and work in harmony. For AI to be used successfully in an organization, that organization needs to have a good balance of computer/tech experts and scientists.

AI is a tool that can accelerate the drug discovery process. It needs to be seen as a tool, not a replacement of the scientist. Likewise, AI will not replace physicians, at least in the foreseeable future. For example, in breast cancer diagnosis, AI will not replace the radiologist who reads mammogram images. Rather, AI will allow the radiologist to do a better job by flagging things to pay special attention to. This is especially important in Countries where there is a shortage of radiologists, such as Mexico. Radiologists are inundated by a massive amount of images – mostly normal - to review, and may miss abnormalities, especially when they are tired after a long day at work.

Physician adoption is a bottleneck for the use of AI in medicine at the present time. Some physicians see the value of a new technology, while others are more skeptical and rely only on what they learn in medical school or at conferences. But a shift is happening, especially in oncology. Surprisingly, older physicians tend to be more open to AI than newly minted ones.

For now, the low hanging fruit in AI are in diagnostic imaging, such as radiology and pathology, but AI can be used more broadly to automate drug discovery processes or select the right patient population for a clinic study.

Retaining intellectual property is not a very important consideration in the AI space. Actually, you need to be careful about what you disclose, because what is in the public domain can be easily copied and slightly changed by competitors. In the software realm, it is better to sell the software and continue to improve it, than to protect it with a patent.

For a chemist or a biologist, it can be challenging to incorporate AI in their company’s business model and AI talent is still in relatively short supply, but things are changing. For example, UCSD has the largest data science institute in the US and AI is rapidly growing. In addition, events like these stimulate discussion and ideas, and may accelerate the incorporation of AI in science and medicine.

The audience was very engaged and the panel discussion was followed by Q&A, after which Rebecca invited everybody to a fantastic reception in the outdoor patio.