We’re putting too much faith in AI’s ability to say no
By Arthur Holland Michel · Oct 9, 2026, 4:00 AM CDT
Ever since people first seriously contemplated giving machines an intelligence modeled on our own, there has never been any question that they would, like us, be able to say no. The sci-fi canon is full of stories of robotic disobedience. Most of these capers are, of course, cautionary. But recently, the idea that AI shouldn’t do everything you ask has become something like a commandment. In 2021,
Excerpt shown under fair-use limits. Full text remains with the original publisher.
People in this coverage
Explore their history and attributable record. Being mentioned does not imply endorsement.
Layer 1 · Claims & fact checks
AI analysisLayer 3 · Reporting analysis
AI analysisAnecdoteUses personal testimony to illustrate early model behavior, lending a human‑scale illustration but not systematic evidence.
MetaphorEmploys vivid metaphor to dramatize the risk, signaling an alarmist tone.
Appeal to AuthorityCites a high‑profile figure to bolster the argument, though the quote is not contextualized with broader expert consensus.
Technical JargonIntroduces specialized terminology to convey scientific rigor, but the underlying study is not fully referenced.
Swiss‑cheese AnalogyUses a familiar analogy to explain layered safety mechanisms, reinforcing the narrative of inherent imperfection.
Context
AI analysisMissing context
The article lacks quantitative data on refusal success rates, comparative analysis of different companies’ safety architectures, and discussion of existing regulatory or policy frameworks governing AI refusal.
Important context
AI safety research includes red‑team testing, classifier layers, and ongoing fine‑tuning; refusal is a core component of current safety strategies, but its effectiveness remains an open research question.
Opinion vs. reporting
AI analysisThe piece blends reporting of interviews and internal practices with strong editorial commentary and speculative warnings; much of the narrative is opinion‑laden.