Marcus believes that, all too often, employers experimenting with replacing employees with AI are underestimating what humans can do and overestimating the technological readiness of current AI, which can often do large portions of a job but not the whole thing.

Not graded on evidence

Outcome: Can’t be checked

An opinion, a recommendation, or too vague to test. Listed, not scored.

Evidence up to 2026-10-28. Checked 2026-10-31.

Why

Marcus frames this as his own belief ('I believe'). He says that employers experimenting with replacing staff with AI 'all too often' underestimate humans and overestimate current AI. The claim is about what employers believe, which cannot be observed directly. 'All too often' sets no frequency threshold. The evidence packet notes that no measure exists of how often employers misjudge human or AI ability. The only evidence item is an Entrepreneur report dated 9 May 2025 that draws on Bloomberg (e1). It says Klarna's CEO now views the company's AI-focused strategy as not the right path, and that its AI chatbots gave 'lower quality' output. That is one employer's account. It fits the view Marcus describes, but it cannot show how often employers in general misjudge AI or humans. So the claim is not scored.

Caveats

• The only evidence is one company's account. It does not show how common the pattern is among employers testing AI replacement.
• The embedded point that current AI 'can often do large portions of a job but not the whole thing' is not assessed on its own.
• The packet did not look at studies or surveys of employer AI adoption or reversals.

Evidence (1 item)

Claims like this aren’t scored, so there is no evidence to show.