I'm excited about the use of AI in improving health worldwide, but concerned that we will be making many mistakes before we get it right, attempting to fly before we can walk. Issues such as sycophancy, poor or non-existent explainability and persistent resort to inventions/"hallucinations" make AI suitable for use only in trained human hands, IMHO.
It is particularly alarming when it is unleashed to offer unsupervised advice in such areas as mental health - and, more broadly, anything to do with psychology (including suicide advice and romance).
Best,
Chris
Chris Zielinski
Centre for Global Health, University of Winchester, UK and President, World Association of Medical Editors (WAME)
Blogs; http://ziggytheblue.wordpress.com and http://ziggytheblue.tumblr.com
Publications: http://www.researchgate.net and https://winchester.academia.edu/ChrisZielinski/
HIFA Profile: Chris Zielinski held senior positions at the World Health Organization for 15 years, in Africa, WHOs Geneva Headquarters, and India, and earlier in other UN-system organizations working in writing, media, publishing, knowledge management, and intellectual property. He also spent three years as Chief Executive of the Authors Licensing and Collecting Society (looking after the intellectual property revenues of all UK authors and journalists). Chris was the founder of the ExtraMED project (Third World biomedical journals on CD-ROM), and managed the Gates Foundation-supported Health Information Centres project. At WHO he was appointed to the Ethical Review Committee, and was an originator of the African Health Observatory during his years in Brazzaville. With interests in the information, and computer ethics and bioethics, Chris has edited numerous books and journals and worked as a translator. Now working independently, Chris has recently finished writing a travel book called Afreekinout. Email: chris AT chriszielinski.com