ABAB Pedia

Karan Singhal

Karan Singhal is an AI researcher who leads the Health AI team at OpenAIOpenAIOpenAI is an artificial intelligence research and product organization founded in 2015 that develops the GPT model series, ChatGPT, Codex and a developer API, and conducts research on image, speech, video and intelligent agents. After its 2025 restructuring, the nonprofit OpenAI Foundation continues to control the for-profit OpenAI Group PBC. Its growth has involved large-scale fundraising and compute partnerships, and has also been accompanied by controversies over governance, copyright, privacy and AI safety.Open the full entry. At Google he worked on medical AI, privacy and safety and helped develop the Med-PaLM series. His research includes how to evaluate language models intended for clinicians and the public.

Contents7 sections
Key facts

Student years at Stanford and teaching AI

Singhal earned his bachelor's and master's degrees in computer science at StanfordStanford UniversityStanford University is a research university in California founded by Leland and Jane Stanford in 1885 and opened in 1891.Open the full entry. He recalled enjoying mathematics and considering physics, medicine, philosophy and economics. As graduation approached he was still deciding between research and product management. He helped establish and teach Stanford's AI for Social Good course on applying computing to social problems.⁠[1][2]

Medical AI research at Google

Over roughly five years at GoogleGoogleGoogle is an American technology company founded by Larry Page and Sergey Brin and incorporated in 1998. It became an Alphabet subsidiary in 2015. Its businesses include internet search, advertising, software platforms, cloud computing, and artificial intelligence.Open the full entry he worked on learning while preserving data privacy, representation learning and AI safety. He dated his team's early discussions of medical language models to late 2020. He helped develop Med-PaLM and Med-PaLM 2 and worked on evaluating how accurately and reliably models handled medical questions with input from physicians. The work appeared in Nature and reached users through Google Cloud.⁠[3][4][2]

Interview with No Priors

He used medical question answering to explain why multiple-choice scores could not characterize long answers, which also needed review for factual accuracy, bias and safety.⁠[5]

OpenAI’s health team and evaluation

Advice Shared on No Priors

He suggested judging a health-information tool against the information people would otherwise have obtained in the same situation. (15:11)⁠[5]

He recommended testing a research model in a specific workflow and assessing both near-term safety and longer-term outcomes. (15:56)⁠[5]

Related Organizations, Websites and Public Accounts

OpenAI: Health AI team lead. Official website: https://openai.com/ (opens in a new window).⁠[4]

Sources7 sources

  1. Minds, Medicine & Machines: AI and the Future of Healthcare (opens in a new window)cfa-sf.org↑
  2. OpenAI’s Karan Singal on HealthBench and the Future of Medical AI (opens in a new window)ai-podcast.nejm.org↑
  3. Karan Singhal (opens in a new window)karansinghal.com↑
  4. Joining OpenAI (opens in a new window)karansinghal.com↑
  5. No Priors Ep. 17 | With Karan Singhal (opens in a new window)youtube.com↑
  6. Notes (opens in a new window)karansinghal.com↑
  7. TIME100 Health: Karan Singhal (opens in a new window)time.com↑