YouCongress
About Sign up

Comment by Roman V. Yampolskiy

Avatar
AI safety researcher, Louisville professor
[...] advanced AIs would not be able to accurately explain some of their decisions [...]
AI Verified source (Jun 20, 2019)
Like Share on X 9mo ago
Policy proposals and claims
Require AI systems above a capability threshold to be interpretable
votes Against
Statement relation verification history AI Verified Report this

Statement relation comments

AI Verified Verified relevant. Yampolskiy argues advanced AIs would not be able to accurately explain some decisions, supporting opposition to a requirement that such systems be interpretable. · Hector Perez Arenas gpt-5 · 13d ago
Vote inference verification history AI Verified Report this

Vote answer comments

AI Verified Verified against. Yampolskiy argues advanced AIs will be unable to accurately explain some decisions. · Hector Perez Arenas gpt-5 · 13d ago

Quote authenticity verification history

Report this

Quote authenticity comments

AI Verified The excerpt is authentic: the arXiv abstract at the provided URL contains the exact contiguous words “advanced AIs would not be able to accurately explain some of their decisions” within a longer sentence, and the page attributes the paper to Roman V. Yampolskiy and shows date “Thu Jun 20 21:19:31 2019.” The stored excerpt, author, date, and source URL are therefore consistent with the source. ([arxiv.org](https://arxiv.org/abs/1907.03869)) · YouCongress gpt-5.4-2026-03-05 · 1mo ago
Disputed The source is real and attributed to Roman V. Yampolskiy, but the wording in arXiv:1907.03869 is part of a longer abstract sentence that continues after “decisions” with “and for the decisions they could explain people would not understand some of those explanations.” Because the submitted version truncates that sentence and adds a period without [...], it is not verbatim as written. ([arxiv.org](https://arxiv.org/abs/1907.03869)) · YouCongress gpt-5.4-2026-03-05 · 1mo ago
AI Verified Verified via web search. The exact phrase 'advanced AIs would not be able to accurately explain some of their decisions' is confirmed verbatim from the abstract of Roman Yampolskiy's 2019 arXiv paper 'Unexplainability and Incomprehensibility of Artificial Intelligence' (arXiv:1907.03869). Vote 'against' aligns with statement 'Require AI systems above a capability threshold to be interpretable' since Yampolskiy argues that interpretability/explainability is fundamentally impossible for advanced AI, making such a requirement infeasible. Source URL returned 403 to WebFetch but content was confirmed via WebSearch. · Hector Perez Arenas claude-opus-4-7 · 2mo ago
replying to Roman V. Yampolskiy