https://m-cacm.acm.org/blogs/blog-cacm/273791-two-models-of-ai-oversightand-how-things-could-go-deeply-wrong/fulltext

Joseph Thornton

	Roxana Daneshjou MD/PhD ⁦‪@RoxanaDaneshjou‬⁩ NEW PAPER ALERT: Automating AI transparency with image-text models! Using dermatology as a use case, we build MONET (Medical cONcept rETriever) that allows to do large scale data auditing, model auditing, and build inherently interpretable models! medrxiv.org/content/10.110… pic.twitter.com/NVRNIoaV5s 6/12/23, 1:44 PM

Joseph Thornton

Joseph Thornton

	Nico Dosenbach ⁦‪@ndosenbach‬⁩ ‘Precision Neuroimaging for Localization-Related Psychiatry’ JAMA Psychiatry opinion w/ Tim Laumann & Chuck Zorumski: ” … mapping the brain’s true functional organization, not the fictitious group-averaged brain, is just beginning.” jamanetwork.com/journals/jamap… pic.twitter.com/K3E0zXEUrZ 6/8/23, 5:35 AM

Joseph Thornton

https://www.neuromodulationjournal.org/article/S1094-7159(23)00448-8/fulltext

Joseph Thornton

	Eric Topol ⁦‪@EricTopol‬⁩ 🆕 ⁦‪@Nature‬⁩ An important LLM #AI for healthcare report using all unstructured clinical notes ⁦‪@nyulangone‬⁩ from 336,000 patients to predict readmission, mortality, comorbidity, and insurance denials nature.com/articles/s4158… by ⁦‪@ekoermann‬⁩ and colleagues pic.twitter.com/83Q1NtDETN 6/7/23, 12:25 PM

Joseph Thornton

	John Torous, MD MBI ⁦‪@JohnTorousMD‬⁩ .⁦‪@sjdalf‬⁩ discuss this digital mental health with topics ranging from novel biomarkers to ethics in this impressive video / panel 6/6/23, 8:51 PM

Joseph Thornton

	John Torous, MD MBI ⁦‪@JohnTorousMD‬⁩ “Risk scores in digital psychiatry: Expanding the reach of complex smartphone data by condensing it into simple results”: Carsten Langholm from our team led this new paper turning multi-dimensional and temporal digital phenotyping into an actionable score: sciencedirect.com/science/articl… pic.twitter.com/MnBImh1l8q 6/5/23, 5:14 PM

Joseph Thornton

	Stanford HAI ⁦‪@StanfordHAI‬⁩ A ⁦‪@Stanford‬⁩ study on AI’s political persuasiveness serves as a cautionary tale for how chatbots could endanger democracy and national security. stanford.io/3OW6cZX 6/5/23, 1:01 PM

Joseph Thornton