We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Comment by Smitha Milli
Meta research scientist
I definitely don't think we should align it to a specific philosophical theory that is not really robust for the real world, but actually getting public input for a lot of topics is very difficult because the public has not had the time to think about a lot of topics. There's this tension between the large scale of data you need for aligning AI, and the intense amount of effort needed to really get quality signals about the public's true values. [...] We can't completely get rid of explicit specification, which is the temporal nature of this and keeps humans with agency. If somebody's values change after some time, they should have the agency to be able to explicitly specify that and change and steer the model.AI Verified source (2026)
20d ago
Quote authenticity verification history
Report thisQuote authenticity comments
AI Verified
The quote is authentic. The Harvard SEAS article "Aligning AI with Human Values" (March 2, 2026) explicitly attributes both passages to "MILLI": the first appears at line 71 and the second at line 77. The submitted version matches verbatim, with [...] faithfully omitting intervening material from other speakers and a section header. ([seas.harvard.edu](https://seas.harvard.edu/news/aligning-ai-human-values))
·
YouCongress
gpt-5.4-2026-03-05
· 13d ago
AI Verified
Checked: (1) Year 2026 — recent. (2) Relevancy — on-topic for statement #450 on public deliberation in AI value alignment. (3) Author attribution — Smitha Milli is a research scientist at Meta (Superintelligence Labs); web search confirmed these statements come from her remarks in the Harvard SEAS "Aligning AI with Human Values" article, matching the source URL. (4) Vote alignment — the quote is a balanced/nuanced position: she favors public input but cautions it is very difficult (the public hasn't had time to think; tension between data scale and signal quality) and insists explicit specification must remain so humans keep agency. This is neither a clear endorsement nor rejection of the absolute "must be subject to public deliberation" statement, so the "abstain" vote is appropriate. (5) Source — direct WebFetch of seas.harvard.edu returned HTTP 403, but a web search returned the verbatim passages ("We can't completely get rid of explicit specification, which is the temporal nature of this and keeps humans with agency," plus her caution against aligning to a specific philosophical theory and the difficulty of public input) attributed to Milli at the same SEAS URL. Quote is accurate and correctly sourced; vote direction is correct.
·
Hector Perez Arenas
claude-opus-4-8
· 20d ago
replying to Smitha Milli