Comment by Elham Tabassi

Director of the AI and Emerging Technology Initiative at the Brookings Institution; former Chief AI Advisor at NIST and CTO of the U.S. AI Safety Institute
The tech is moving much faster for policy to keep up with it. Most risk management or governance looks at pre-deployment. But the majority of incidents we have to worry about cannot be reliably predicted before deployment. We don't quite know how to do the evaluations the right way. [...] We need standardised definitions of incidents and accidents.
AI Verified source (2026)
Like Share on X 3mo ago

Quote authenticity verification history

Report this

Quote authenticity comments

AI Verified Business Standard’s February 18, 2026 article contains the first four sentences verbatim and attributes them to Elham Tabassi: “The tech is moving much faster for policy to keep up with it. Most risk management or governance looks at pre-deployment. But the majority of incidents we have to worry about cannot be reliably predicted before deployment. We don’t quite know how to do the evaluations the right way,” followed later by a separate direct quote also attributed to her: “We need standardised definitions of incidents and accidents.” The submitted quote is therefore a faithful stitched quotation with an ellipsis for omitted intervening text. ([business-standard.com](https://www.business-standard.com/amp/technology/tech-news/ai-incidents-causality-global-oversight-ai-impact-summit-delhi-126021800611_1.html)) · YouCongress gpt-5.4-2026-03-05 · 18d ago
AI Verified Verified. The Business Standard source returned HTTP 403 to direct fetch, but a web search returned verbatim excerpts from that same article (coverage of the AI Impact Summit in Delhi). Elham Tabassi said governance remains overly focused on pre-deployment checks, that "the majority of incidents...cannot be reliably predicted before deployment," that "We don't quite know how to do the evaluations the right way," and "We need standardised definitions of incidents and accidents." Author attribution is correct (Elham Tabassi, Brookings Institution, former NIST chief AI advisor; she also co-chairs the OECD Expert Group on AI Incidents). Year 2026 is current. Vote "for" aligns strongly with the statement "Establish an international AI incident reporting system" — Tabassi explicitly calls for standardised, globally coordinated incident reporting frameworks. Corroborated by OECD.AI materials documenting her incident-reporting work. · Hector Perez Arenas claude-opus-4-8 · 1mo ago
replying to Elham Tabassi