Read on Twitter
I've written a short piece on the #spiritAI & #consortAI extensions published today.
https://ada.com/editorial/what-do-the-new-consort-and-spirit-guidelines-mean-for-health-ai/
I don't want to overestimate my role—I was but a small cog in a machine very ably run by @DrXiaoLiu @samsamcr & others—but had some thoughts on their wider context 1/7 https://twitter.com/bmj_latest/status/1303711908206440448
https://ada.com/editorial/what-do-the-new-consort-and-spirit-guidelines-mean-for-health-ai/
I don't want to overestimate my role—I was but a small cog in a machine very ably run by @DrXiaoLiu @samsamcr & others—but had some thoughts on their wider context 1/7 https://twitter.com/bmj_latest/status/1303711908206440448
Firstly, it's extremely exciting to see clinical #AI research being brought into the mainstream. If, as a field, we want AI adoption to be seriously considered by healthcare professionals and systems, then we need to meet the high standards set for all clinical research. 2/7
In fact, this should be the foundation on which successful AI adoption is built—a quest for high medical quality, underpinned by strong evidence resulting from robust and transparent studies. 3/7
I see these extensions as part of wider shift going on in field—'operationalising' development of healthcare AI. Deciding *what* we should do about these novel technologies is easy; *how* we should do it is much harder. @DrXiaoLiu & colleagues have grasped that nettle. 4/7
Also, the time for hype around the use of AI in healthcare is over—efforts like these force the field to be honest with ourselves about what AI can and cannot do, and the specific circumstances where & when it can do this. 5/7
For some more perspectives on this, take a look at @PaulLikeMe, @DrXiaoLiu & @Denniston_Ophth's editorial (with what I guarantee is the best title you'll read this week): https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-020-01754-z 6/7
And for the final word on all this, for the view on what this means to the people who matter most in all our work, read @MannaElaine's piece: https://www.nature.com/articles/s41591-020-1053-7 7/7
