This analysis describes what OpenAI's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
This requirement places the burden of verification and professional judgment on the user, limiting reliance on the output for clinical decision-making.
Interpretive note: The excerpt uses 'You should always' rather than 'You must' or 'You are prohibited from', which may indicate a strong recommendation rather than a binding obligation. The canonical claim uses 'requires' consistent with the directive framing, but the precise legal force of 'should' versus 'must' is ambiguous without broader contract context.
The updated terms introduce a new Prompt-to-Voice feature with strict usage requirements. Users must obtain express consent before reproducing any person's voice, must clearly disclose that generated speech is AI-generated, and must not remove or alter OpenAI-provided AI-origin indicators. Critically, OpenAI's indemnification obligations do not apply to Prompt-to-Voice or its output, meaning customers and End Users bear full legal liability if generated speech infringes third-party intellectual property rights or violates consent requirements.
View change record →The updated terms establish new responsibilities for users who deploy autonomous agents (Dots). Users must provide clear instructions, set appropriate limits, monitor agent actions, and are liable for transactions the agents make on their behalf. Additionally, users connecting their ChatGPT accounts to third-party services via Sign in with ChatGPT will share authentication information and usage data with those services, subject to those services' own terms and policies. You can review the Sign in with ChatGPT Terms and the third-party service's privacy policy before connecting your account.
View change record →The updated terms introduce OpenAI Academy, a free learning platform, and establish separate governance for it. If you access Academy through an organization-associated email address or account (including via Sign in with ChatGPT with a workspace account), OpenAI states it may share your Academy activity data—including enrollment, engagement, progress, and completion status—with your organization's owners or administrators. Business or enterprise workspace terms do not apply to Academy use, even when you sign in with a workspace account. You can review the separate OpenAI Academy Terms of Use and Code of Conduct to understand the specific rules governing Academy access.
View change record →If you use ChatGPT for Healthcare, you are required to independently verify its output and exercise professional judgment rather than relying primarily or solely on it for patient decisions.
How other platforms handle this
We kindly request you do not disclose certain Data, such as passwords, credit card numbers, PINs or other confidential data during your use of our Services.
You may not use our Services in any manner that infringes the rights of any third party.
If you disable or delete cookies, or if you are running third-party software that intercepts or deletes cookies, please note that some parts or functionality of our Websites may not work properly.
"You should always verify the information provided by ChatGPT for Healthcare and exercise independent professional judgment in decision-making about a patient without relying primarily or solely on the output.Excerpt from OpenAI's Service Terms
Get the weekly coverage letter
Companies change their terms without announcing it. We capture governance documents, compare each new version with the last, and publish what changed, with the source text behind it. One email a week on what we found and what we could not reach.
This requirement places the burden of verification and professional judgment on the user, limiting reliance on the output for clinical decision-making.
If you use ChatGPT for Healthcare, you are required to independently verify its output and exercise professional judgment rather than relying primarily or solely on it for patient decisions.
ConductAtlas has identified this type of provision across 272 platforms. See the full comparison.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by OpenAI.