Discussion

• @marypcbuk.bsky.social Mary Branscombe on bluesky

  Google wants to know a lot about everyone's health and activity because they can do more with your data than you can by keeping it private, how selfish of you not to be keen on that [embedded post]

• @google @google on x

  Access to straightforward, evidence-based information helps people better navigate their health. Learn more about our latest updates announced at @GoogleHealth's Check Up event ↓ https://blog.google/...

• @amycchou Amy C. Chou on x

  Check out our team's latest paper, “Health equity assessment of machine learning performance (HEAL): a framework and dermatology AI model case study."" Congratulations to all involved! https://www.thelancet.com/...

• @katieelink Katie Link on x

  Looking forward to diving into this preprint on an important subject! Btw, I quickly uploaded the EquityMedQA CSVs to @huggingface in case anyone else is interested in taking a look at them in your browser (rather than downloading from arxiv): https://huggingface.co/... [image]

• @ashleycapoot Ashley Capoot on x

  Caught up w/ @GoogleHealth at #TheCheckUp this morning! The company shared updates about its ongoing research/use of AI in health care. There was a big focus on early-stage projects and initiatives. Here are a few things that stuck out to me: [image]

• @berci @berci on x

  What an exemplary achievement from Google not only to define how health equity differs from equality, but to also provide a framework to improve the former. https://blog.research.google/ ...

• @stephenpfohl Stephen Pfohl on x

  Excited to share new work on surfacing health equity-related biases in LLMs. We design rubrics covering 6 dimensions of bias, release EquityMedQA, a collection of 7 adversarial datasets, and conduct a large-scale human evaluation study with Med-PaLM 2. https://arxiv.org/... [imag…

• @liamgmccoy @liamgmccoy on x

  A pleasure to collab on this project with @GoogleHealth researchers. Bias and equity risks in LLMs are complex, given the open-ended nature of questions and responses. This large scale human assessment framework is an important step forward: We cannot address what we do not see!

• @hcolelewis Heather Cole-Lewis on x

  Grateful to introduce our new paper “A Toolbox for Surfacing Health Equity Harms and Biases in Large Language Models”. Adversarial datasets and human assessment rubrics to evaluate LLMs for potential biases and equity-related harms: https://arxiv.org/... 🧵: [image]

• @ashleycapoot Ashley Capoot on x

  In her opening remarks, Google's @KBDeSalvo said “it seems clear that in the future AI won't replace doctors, but doctors who use AI will replace those who don't.” Pretty clear where the co thinks health tech is going. (My transcript says AI was mentioned 78 times at the event!)

• @googlehealth @googlehealth on x

  In the next ten years, @HospitalsApollo will use our AI models to provide free tuberculosis, lung cancer and breast cancer - helping thousands more people across India access care sooner. More from our latest #TheCheckUp announcement: https://blog.google/...

• @googlehealth @googlehealth on x

  Google's new Personal Health LLM will bring personalized coaching and other health and wellness experiences to our users. At #TheCheckUp, we shared how Fitbit and Google Research are partnering with experts to develop (cont) http://spklr.io/...

• @googleorg @googleorg on x

  With support from https://google.org/, @armmanindia is tackling maternal and child health issues with mMitra: a free, AI powered service that sends timely and targeted phone messages to new and expectant mothers. To date, mMitra has reached 3 million women across India.

• @thekaransinghal Karan Singhal on x

  Excited to share our newest work! 📝 Evaluation of LLMs is hard, especially for health equity. We provide a multifaceted human assessment framework, 7 newly-released adversarial datasets, and perform the largest human eval study on this topic to date. 🧵: https://arxiv.org/... [ima…