Deb Raji
@rajiinio.bsky.social
AI accountability, audits & eval. Keen on participation & practical outcomes. CS PhDing @UCBerkeley.
More of this kind of reporting, please!
September 12, 2025 at 5:11 PM
More of this kind of reporting, please!
Was beyond disappointed to see this in the AI Action Plan. Messing with the NIST RMF (which many private & public institutions currently rely on) feels like a cheap shot
July 24, 2025 at 2:25 PM
Was beyond disappointed to see this in the AI Action Plan. Messing with the NIST RMF (which many private & public institutions currently rely on) feels like a cheap shot
July 1, 2025 at 3:19 PM
So much of the tech hubris about "revolutionizing" government is an assumption that, just because these orgs work so differently from their own, they must be doing something "wrong".
It's an insulting level of under-appreciation for very good & stable systems that just.. aren't a tech company.
It's an insulting level of under-appreciation for very good & stable systems that just.. aren't a tech company.
June 4, 2025 at 6:24 PM
So much of the tech hubris about "revolutionizing" government is an assumption that, just because these orgs work so differently from their own, they must be doing something "wrong".
It's an insulting level of under-appreciation for very good & stable systems that just.. aren't a tech company.
It's an insulting level of under-appreciation for very good & stable systems that just.. aren't a tech company.
I'll be at Stanford later today, to give a talk on ML evaluation!
For those around & interested, please drop by:
For those around & interested, please drop by:
May 8, 2025 at 2:19 PM
I'll be at Stanford later today, to give a talk on ML evaluation!
For those around & interested, please drop by:
For those around & interested, please drop by:
Forgot to share this work here as well:
In medical settings, researchers & institutional users often cite medical examination benchmark results as a proxy for AI model performance.
This is a woefully misguided practice - as with actual doctors, exam performance does NOT equal clinical competence!
In medical settings, researchers & institutional users often cite medical examination benchmark results as a proxy for AI model performance.
This is a woefully misguided practice - as with actual doctors, exam performance does NOT equal clinical competence!
January 28, 2025 at 5:47 PM
Forgot to share this work here as well:
In medical settings, researchers & institutional users often cite medical examination benchmark results as a proxy for AI model performance.
This is a woefully misguided practice - as with actual doctors, exam performance does NOT equal clinical competence!
In medical settings, researchers & institutional users often cite medical examination benchmark results as a proxy for AI model performance.
This is a woefully misguided practice - as with actual doctors, exam performance does NOT equal clinical competence!
Really wonder what they're pitching here lol
January 20, 2025 at 11:11 PM
Really wonder what they're pitching here lol
Whoa - great to see this work cited in the bipartisan House report on AI: obernolte.house.gov/AITFReport
December 18, 2024 at 6:55 AM
Whoa - great to see this work cited in the bipartisan House report on AI: obernolte.house.gov/AITFReport