Let The Public Decide On AI For Films

Hollywood is up in arms at the use of AI in films, with encroachment from all sides, such as AI actors like Tilly Norwood (with “40 more to come” says its creator) and short AI films by Sora and Seedance. But not to worry! Even if AI ultimately creates perfectly crafted feature-length films that are indistinguishable from non-AI films, the public might not embrace them – if they know.
The moment I realize that a short is created by AI (as is now common), I flip to the next one, since simply knowing it is created by AI drains out all the fun and the drama, even if the short appears perfectly realistic. Just knowing that it is created by AI leads to a shoulder shrug “next” attitude.
Imagine watching the drama of the World Series, and then being told in the middle of it that it is all created by AI, with no real players. You would almost certainly turn it off, even if it were frame-by-frame identical to the actual games. Who cares to watch if all the action and “drama” is created by AI? Yawn…
It will be the same with feature-length films. Just knowing that they are created with AI may deflate attendance. Part of the success of the Mission Impossible films is knowing that Tom Cruise does the stunts himself. Imagine the lack of enthusiasm if the public thought the entire film was created by AI.
Indeed, a selling point of films in the future may be that films do not have any AI elements, particularly if films with AI elements become common. All that needs to happen is to give fair warning to the public when a film uses AI, and the law is marching in that direction:
1. Many states require the disclosure of the use of AI in political advertisements, and New York has expanded this disclosure to be required in almost all advertisements (whether or not political) that include an image of a human created by AI. It would be a short leap to extend that to films.
2. The EU requires the disclosure of the use of AI in all films starting in August, 2026. The language is ambiguous, but disclosure in the end credits will probably suffice. The problem with that approach is that the public should know before they watch the film, so the disclosure should come at the start of the film, such as in the main titles.
3. There is pending U.S. federal legislation (the AI Labeling Act) that would require “clear and conspicuous disclosure” of the use of AI-generated content in almost all content, including films. The requirement of “clear and conspicuous disclosure” probably requires disclosure in the main titles, and the Act requires audible disclosure if there is audible AI content in the film.

If the MPA (formerly the MPAA) wants to head off federally mandated legislation on this issue, including the requirement of an audible warning, then it should add an “AI” rating for films that use any AI, just as it used the current rating system to head off federally-mandated censorship in the 1960’s. Under this approach, a distributor would certify that AI was or was not used in some manner in connection with the film (including the screenplay, actors, music, or scenery), and if AI were used, the film would receive an “AI” rating next to the regular rating (e.g., “PG” or “R”). In this manner, the public is at least given fair warning, and from there the market can decide.