Cas (Stephen Casper)
@stephenlcasper
AI technical governance & risk management research. PhD Candidate @MIT_CSAIL / @MITEECS. Also at scasper.bsky.social.
stephencasper.com
ID: 704559922143322112
http://stephencasper.com 01-03-2016 06:52:30
1,1K Tweet
4,4K Followers
3,3K Following
![📣 New paper
AI gov. frameworks are being designed to rely on rigorous assessments of capabilities & risks. But risk evals are [still] pretty bad – they regularly fail to find overtly harmful behaviors that surface post-deployment.
Model tampering attacks can help with this.](https://pbs.twimg.com/media/Gjg3cH0aIAAWM2f.jpg "Cas (Stephen Casper) (@stephenlcasper) on Twitter photo 📣 New paper
AI gov. frameworks are being designed to rely on rigorous assessments of capabilities & risks. But risk evals are [still] pretty bad – they regularly fail to find overtly harmful behaviors that surface post-deployment.
Model tampering attacks can help with this.")