AI and Fact-Check Disclaimer

Effective [Effective date]

Verity uses AI and automated tools. Here's what its results are, and what they aren't.

Fact checks

• Verity searches the web, keeps only sources it treats as trusted (such as government agencies, major news organizations, and established references), and asks an AI to judge the claim against what those sources say. It labels claims True, False, Mixed, or Can't verify.
• Results can be wrong. Speech recognition can mishear words, the AI can misread a claim or a source, sources can be outdated or wrong, and new facts may not be online yet.
• Always open the linked source before relying on a result. If there's no good source, Verity says it can't verify rather than guess.
• Opinions and predictions aren't fact-checked.

Contradictions, dodged questions, and interruptions

• These flag statements, not people. A flag is a prompt to look closer, not proof that someone lied, evaded, or was rude.
• Each flag shows the exact words and times so you can judge for yourself, and you can dismiss any you disagree with.
• Interruptions are estimated from timing and sentences that were cut short, because transcription doesn't capture two people talking at once.
• Speaker detection is automatic and can mix people up.

Verity Assist

Rephrasings aim to keep your meaning, but can change nuance. Read or listen before you use them, and don't use them to make claims you don't mean.

Not professional advice

Nothing in Verity is legal, medical, financial, or other professional advice. Don't rely on it for decisions where mistakes could cause harm.