Yoshua Bengio @Yoshua_Bengio · 03 Feb 2026
Because no single safeguard reliably prevents misuse or malfunctions, developers are converging on "defence-in-depth." This means layering multiple measures—model-level training, input/output filters, monitoring, access controls, and governance—so that if one fails, others may https://t.co/CSnO4ypRgx
4 053Views
44Likes
7Reposts
2Replies
0Quotes
6Bookmarks
Is that a lot?
2.08×vs 10K–100K median1 948 views is typical
8.94%Reachviews ÷ followers
1.31%Engagement rateof viewers reacted