Comment by Sam Altman

When the frontier models are in the hands of pretty responsible companies, I think we can mitigate [risks] by the companies aligning their models and having good classifiers and good safety stacks. Unverified source (2026)