The recent excitement has been ignited by the release of AI-powered technologies, with the most high profile being ChatGPT, an AI chatbot trained to follow a conversational instruction and provide a detailed response. ChatGPT can answer follow up questions, admit its mistakes, challenge incorrect premises and reject inappropriate requests. It was developed by OpenAI and released in November.
Adrian Lu, an investment analyst at Magellan Financial Group, describes AI as computers capable of thinking and understanding the world around them – that is, they can reason, learn and act with autonomy. He believes the pinnacle of AI is likely to be far away, but engineers are getting better at building models that can mimic human perception, behaviour and abilities. These models can already perform tasks better and faster than humans, which promises vast productivity gains as well as potential threats to investors.
Platinum Asset Management chief executive Andrew Clifford says investors are excited that AI could be a major disruptive force in the global economy. It reminds him a lot of the excitement around data on mobile phones in 1999 when NTT DoCoMo first pioneered full internet access on a mobile phone in Japan.
It took around five years for that to translate into widespread economic outcomes and significant revenue for a company such as Research in Motion (maker of Blackberry) which was then decimated by the release of Apple’s touchscreen iPhone in 2007. With that in mind, Clifford says AI could be a genuine investible theme, but investors need to be careful in this initial period.
Investors have been quick to identify many of the primary beneficiaries of the ongoing development of AI. Lu characterises these as the “enablers” of AI and highlights that many technologies had to come together to make AI possible, from semiconductors to software to hyperscale data centres.
Some leading companies in these enabling technologies have been the greatest beneficiaries of the acceleration in AI spending, including Microsoft in enterprise cloud computing, Nvidia in AI accelerator chips, ASML in chip making equipment and TSMC in leading-edge manufacturing. Investors now need to ask whether these companies’ share prices represent an opportunity or a bubble.
Other beneficiaries from AI – ranging from industrial automation to consumer devices, automotive and healthcare – are yet to be fully appreciated by the market according to Bianca Ogden, portfolio manager at Platinum Asset Management. In the healthcare sector, Ogden points to research and drug discovery.
AI is used to assist and ultimately design new therapies (small molecules and biologics) that have the desired attributes. The issue today is that the process from target to lead compound takes roughly four to five years and requires testing of a large number of molecules that then have to be refined over and over again. AI is showing great promise to reduce that timeline.
Oxford-based Exscientia is one of the companies leading the way and works closely with Sanofi, a company that has put together an impressive network of AI partners. Vancouver-based Absci is another interesting company working closely with Nvidia to make the discovery of antibody therapeutics more efficient using its own wet-lab generated database along with AI tools.
Lu highlights some risk to investors with AI, including political and regulatory ones that will touch companies in different ways, and not just the AI enablers. Intellectual property ownership, misinformation, data privacy and jobs displacement are among the key issues.
Clifford says that from an investment standpoint, the risks he is most mindful of are valuation risk (paying too much for an investment) and technological disruption. Remember Kodak, Blockbuster Video and Blackberry, whose businesses were decimated by new technology. It is very likely that AI will have that impact on various businesses over the coming years.
Technology moves quickly and investors need to balance their thinking between opportunities and the threats from AI.