ClaimsDaniela Amodei

“Sometimes Claude gets it wrong.”

Daniela Amodei, Anthropic president 9 May 2026 Daniela Amodei, Co-Founder and President of Anthropic: Building AI the Right Way (transcript) (The Transcript Desk)

Factual AI capabilities

Claude sometimes gives incorrect answers.

Resolved true The record shows it happened or was so.

Factual: Statements about the present or the past, checked against the record: did they hold up?

The evidence is the person's or their organisation's own records.

Said for an organisation they lead or work for; it counts on both records.

What would count as true or false

True if documented cases, benchmark error rates or Anthropic's own documentation from on or before May 2026 show that Claude sometimes gives incorrect answers. Mostly true if this holds with only a small nuance. Mostly false or false only if credible evidence showed that Claude essentially never errs, which would contradict the claim.

Set before the research began.

Evidence (3)

  1. E1 Claude is providing incorrect or misleading responses. What’s going on? | Claude Help Center Anthropic (Claude Help Center) · 16 March 2026 Own record

    “In an attempt to be a helpful assistant, Claude can occasionally produce responses that are incorrect or misleading.”

  2. E2 Model Card and Evaluations for Claude Models Anthropic · 19 July 2025 Own record

    “Claude models still confabulate – getting facts wrong, hallucinating details, and filling in gaps in knowledge with fabrication.”

  3. E3 Model Card and Evaluations for Claude Models Anthropic · 19 July 2025 Own record

    “Claude 2.1 demonstrated a 30% reduction in incorrect answers and a 3-4x lower rate of mistakenly concluding a document supports a particular claim.”

Reasoning

The resolution criteria accept Anthropic's own documentation from on or before May 2026. Anthropic's Help Center article of March 2026 says Claude can occasionally produce incorrect responses [E1]. Anthropic's model card (E2, E3, July 2025) says Claude models "still confabulate – getting facts wrong." It also reports a 30% reduction in incorrect answers for Claude 2.1, which means the earlier model gave incorrect answers at some rate. None of the evidence suggests that Claude essentially never errs.

Cite this claim

Daniela Amodei, “Claude sometimes gives incorrect answers.” (said 9 May 2026). Verdict: resolved true. Reliability API, claim c_4fadfa25f0fd, method 0.1.0+28da144b, dataset release v1.1.2. https://epistacklabs.com/claims/c_4fadfa25f0fd

This claim is in dataset release v1.1.2, with the same verdict.

Reply to this claim

For Daniela Amodei, or someone speaking for them: a correction or counter-evidence. Replies are read before anything is shown; an approved reply shows your name, role, date, statement and link. Your email address is never shown and is deleted 30 days after the decision (privacy).

More factual claims by Daniela Amodei

  1. Mostly true Factual 21 October 2026
    “In the lab, Claude designed new protein binders from scratch.”

    In laboratory work, Anthropic's Claude designed new protein binders from scratch.

  2. Mostly true Factual 21 October 2026
    “including open-source code much of the internet runs on, before attackers could exploit them.”

    Cyberdefenders directed Claude to find more than 33,000 critical and high-severity flaws in widely used software, including open-source code, before attackers could exploit them.

  3. Resolved true Factual 21 October 2026
    “They also showed that more than 6,000 women had come in for prenatal care, but only 1,762 had delivered with a healthcare worker present, a discrepancy that the team is investigating and addressing.”

    The dashboards showed more than 6,000 women came for prenatal care but only 1,762 delivered with a healthcare worker present.

  4. Resolved true Factual 21 October 2026
    “Those dashboards flagged a spike of 25,000 pneumonia cases, so the team can anticipate and plan for future surges.”

    Claude-built dashboards for an International Rescue Committee team in Burkina Faso flagged a spike of 25,000 pneumonia cases.

  5. Resolved true Factual 29 June 2026
    “Sometimes Claude is wrong.”

    Anthropic's Claude sometimes gives wrong answers.

All of them

Verdicts are made automatically; how verdicts are made.