AI can revolutionise medical research, but a cautious approach is needed

PRI ESPL INT .MELBOURNE TGA1 AI-MEDICAL RESEARCH AI promises to revolutionise medical research, but a cautious approach is needed By Reece Hooker, Assistant Producer, 360info Asia-Pacific Melbourne, Aug 1 (360info) One week social-media users are generating and sharing amusing images using AI, the next we hear AI can predict the structure of over 200 million proteins. AI is particularly well placed to revolutionise medical research. The technology helps in two vital ways: it optimises research, and it can make discoveries that humans have not. AI is unshackled from the limitations that come with human researchers: it can trawl deep datasets exponentially faster, never needs to take a break, and never succumbs to illness or fatigue. Taiwanese computer scientist, businessman and author Kai-Fu Lee recently spoke about his optimism for the future of AI. We've surprised ourselves with how well machine learning algorithms work. It makes us focus on the things that AI cannot do, and it will probably either lead to a greater understanding of the human mystique of how we think or it will lead to more breakthroughs, leading to superintelligence It took 40 years, but I think we're basically there. None of this means the rise of the machines is imminent. Despite its advances, AI is still encumbered by technical limitations. AI pioneer Yosua Bengio said in 2021 that deep learning was not anywhere close today to the level of intelligence of a two-year-old child. A study by Pugliese et al. (2021) finds that more and more medical research is mentioning "machine learning". But maybe we have algorithms that are equivalent to lower animals, for perception. And we're gradually climbing this ladder in terms of tools that allow an entity to explore its environment, he said. Much more work needs to be done to perfect the technology, and there is no foreseeable date by which humans will be superfluous to the world of medical research. There is also the matter of ethics in AI, in which legislators will have to parse a minefield of questions about issues ranging from data collection to intellectual property and beyond. In the meantime, leading minds continue to use machine learning and AI to innovate and expand the horizon of what's possible in medical research. REALITY CHECK Between 2000 and 2019, the research community became much more interested in AI: the global output of AI research grew from 52,000 journal publications and conference papers in 2000 to roughly 403,000 in 2019. AI is projected to contribute US$15.7 trillion to the 2030 global economy. Nine in 10 of the 100 healthcare executives surveyed in 2020 said their hospital had an AI and automation strategy in place. BIG IDEAS "The goal and expectation is that in fully integrating AI into research about disease development, many unknowns will become known. What causes diseases to spread, worsen and change, and additional early signs or symptoms that research has yet to uncover, should reveal themselves with the new frontiers made possible by automated AI," Hiroaki Kitano, Okinawa Institute of Science and Technology. "The use of AI in making medical decisions is still new and many barriers need to be overcome before it is used widely in clinical practice. For it to reach its full capacity, wider research and more rigorous approaches are needed to grapple with the ethical issues it raises. This is an ideal time for medical professionals, stakeholders and governments, as well as individuals and their families, to work together and seek a balance between the benefits and risks of the new technologies," Alexander Merkin, Auckland University of Technology. (360info) VM 08011017 NNNN