New York FAIR News Act: AI video labels and what is pending

The New York FAIR News Act is not law yet. The legislature passed it in June 2026, and it would require news content substantially created by generative AI to carry a conspicuous label at the top of the page, image or video, with a spoken announcement at the start of audio. As of the sources I read it was waiting for Governor Hochul's decision.

The facts come from the New York Senate press release, Ethics and Journalism and the AI Commission (all read 2026-10-02). The last is dated 1 October 2026 and still says the bill awaits the governor, so I could not confirm a signature. Check the state's bill page before you rely on any date.

What would the bill require?

Per the press release, it covers news organizations publishing in New York and mandates clear disclaimers on content substantially or wholly generated by AI. The bill number is S.8451-B / A.8962-B. Ethics and Journalism describes the trigger as content 'substantially composed, authored, or otherwise created' by AI and notes the bill does not fix label wording.

The AI Commission piece lists the enforcer and the money: the New York attorney general, with civil fines of 1,000 dollars for a first violation and 5,000 dollars for later ones, plus injunctions. It also notes an exemption where the content qualifies for copyright registration, and that 'substantially' is undefined.

Does it reach a brand video or an ad?

The sources frame it around news media content and newsrooms. A brand explainer is not obviously 'news', so I would not assume it applies, and the sources give no definition that settles the edge. A branded news-style video that you publish as news would be the cautious case.

Separate rules govern advertising with synthetic performers, which is a different New York law; see our New York synthetic performer ad law post.

How would a label be added to a Sume video?

A label 'at the top' is a burned-in element. Video captions burns authored cues with text, start and end, and Timeline compose can hold a prepared still on screen for the whole video layer. Both produce a new MP4.

Sume does not generate legally approved wording, does not say whether your piece is 'substantially' AI-made, and does not add a spoken announcement for audio as a documented step. If you need the spoken line, record or generate it yourself and place it first in the audio. Sume's docs describe what each tool returns and say nothing about a watermark or embedded provenance record on outputs, so treat the output as unmarked until you have checked the delivered file yourself.

What are the open questions?

The AI Commission piece raises the two that matter most. First, 'substantially' is never defined, so a newsroom cannot tell whether an AI-written draft that a person rewrote, or an AI-generated B-roll shot inside a human-made story, crosses the line. Second, the copyright-registration exemption is confusing, because AI-assisted works can be registrable when they show enough human authorship.

Free speech groups and media industry bodies have urged a veto on First Amendment grounds, per Ethics and Journalism, and the governor has until 31 December to act. So the possible outcomes are a signature with a 60-day runway, a veto, or an amended bill.

For planning, the sensible position is to build the label capability but not to announce compliance with a law that does not yet exist. Treat any claim of an effective date with suspicion until the signed text is public.

What to do while it is pending

Use this as a working list and adjust it to your own pipeline and counsel's advice.

• Do not treat the bill as law until you see a signed chapter.
• Decide your own 'substantially' threshold and write it down.
• Build the label step now; it is cheap and reusable for other rules.
• Keep generation records for news-style clips.
• Re-check status before 31 December 2026.