Anthropic's initial constitution for its models drew on broadly agreed sources such as the UN Declaration of Human Rights and parts of Apple's Terms of Service.
Mostly true True in substance, wrong in a detail.
Factual: Statements about the present or the past, checked against the record: did they hold up?
The second judge would not resolve this claim; the verdict shown is the first judge's.
Said for an organisation they lead or work for; it counts on both records.
What would count as true or false
Judge as of 2023-08-08. True: Anthropic's published constitution for Claude (the May 2023 'Claude's Constitution' post or the Constitutional AI materials) lists the UN Universal Declaration of Human Rights and principles inspired by Apple's Terms of Service among its sources. Mostly true: both sources are cited but described a little differently (e.g. 'inspired by' rather than taken from), or they were part of an early published version rather than the very first internal one. Mixed: only one of the two named sources is documented. Mostly false: the documented sources differ and the named ones played at most a marginal or unclear role. False: Anthropic's documentation shows neither source was used.
Set before the research began.
Evidence (2)
-
E1 Anthropic is all in on 'AI safety'—and that's helping the $183 billion startup win over big business | Fortune Fortune · 2 December 2025
“Anthropic drew the principles for Claude’s current constitution from sources as varied as the UN Universal Declaration of Human Rights, Apple’s terms of service, and rules that Anthropic competitor Google DeepMind developed and published for Swallow, a chatbot it created in 2022.”
-
E2 Anthropic is all in on 'AI safety'—and that's helping the $183 billion startup win over big business | Fortune Fortune · 2 December 2025
“Some of Claude’s better performance is down to a technique Anthropic pioneered called “constitutional AI.” This involves giving Claude a written constitution—a set of principles—that is used to train the model.”
Reasoning
E1 is an independent Fortune report. It says Anthropic drew the principles for Claude's constitution from sources including the UN Universal Declaration of Human Rights and Apple's terms of service, which matches both sources named in the claim. E2 confirms that this constitution is the written set of principles used to train Claude. However, E1 is a secondary report from December 2025 that describes Claude's 'current' constitution. It does not quote Anthropic's May 2023 publication or show that this was the initial version, so the timing match is inferred rather than documented.
Cite this claim
Dario Amodei, “Anthropic's initial constitution for its models drew on broadly agreed sources such as the UN Declaration of Human Rights and parts of Apple's Terms of Service.” (said 8 August 2023). Verdict: mostly true. Reliability API, claim c_b2ec69b08cf7, method 0.1.0+c876030c, dataset release v1.1.2. https://epistacklabs.com/claims/c_b2ec69b08cf7
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More factual claims by Dario Amodei
-
Mostly true
“So we did this experiment where we basically train the model to be all the good thing, helpful, honest, harmless, friendly.”
Anthropic ran an experiment training a model to be helpful, honest, harmless and friendly.
-
Resolved true
“They're not yet at the stage where we think that there is a real and meaningful increase in the threat end to end,”
-
Resolved true
“10 to 25% referred to the chance of civilization getting substantially derailed, right?”
-
Mostly true
“this lack of understanding is essentially unprecedented in the history of technology.”
-
Mostly true
“We were able to provide a strong theoretical basis for having conviction that superposition was the right hypothesis to pursue.”
Verdicts are made automatically; how verdicts are made.