Standards in this industry almost never arrive from the top. They arrive from whoever gets tired of the argument first.
The Synthetic Motion Picture Ratings System went live this week at smprs.org, offering film productions a certification that discloses how much of the work was done by artificial intelligence. Screamify adopted it the same day, making the horror streamer the first platform to commit to rating its own catalog. Every Screamify Original film and television show will carry an SMP rating.
The scale is the substance of it. Six tiers, each with a defined percentage band. SMP-0 Practical-Only permits no AI at all — live actors, real sets, a human script, traditional post — and the system positions it as the most prestigious certification available, the gold standard. SMP-1 Assisted Practical holds AI under 15 percent and restricts it to non-creative cleanup. SMP-2 Enhanced Practical, 15 to 35 percent, admits digi-doubles and AI backdrops around a live cast. SMP-3 Augmented Hybrid, 35 to 60 percent, is the first tier to concede creative authorship, allowing a script co-written up to half by AI. SMP-4 Synthetic-Lean Hybrid, 60 to 85 percent, keeps human principals in majority-synthetic environments. SMP-5 Fully Synthetic, 85 to 100 percent, describes a production with digital performers, AI voices, AI sets and a machine-written script and storyboards.
What distinguishes this from the various AI-labeling proposals that have circulated since 2024 is that it is not a warning. The framing throughout is certification: a filmmaker earns a rating and receives a badge set to carry into distribution. The system treats a low number as an asset to advertise rather than a compliance box, which is a shrewd read of how the market is actually behaving. Practical effects have become a selling point again, and SMP-0 gives that instinct a stamp.
The parallel is less the MPAA than the technical certifications that studios pursue voluntarily because exhibitors and buyers have learned to look for them. Those work through adoption rather than authority. Nobody is compelled to seek one; the value accrues because enough of the supply chain treats the absence of one as a question.
Which makes Screamify's participation the more consequential half of the week's news. A standard with no adopters is a website. The platform has committed to rating its entire Originals slate, and the first title with a public number is not a flattering one for the technology skeptics or the technology boosters. Multiverse, an interdimensional-cable channel arriving in October, is fully AI-generated and is going out rated SMP-5. A platform that intended to use the system as cover would not have led with its most synthetic property.
The unanswered question is verification. Certification is initiated by the production. A registrant supplies the title and details, attests that the information provided is accurate and that they hold the authority to submit on the production's behalf, and pays $29.99 for the rating. Higher-priced tiers advertise enhanced AI-usage verification and distribution-ready documentation, but the baseline mechanism is self-declaration.
That is not disqualifying. Self-declaration is where most voluntary standards begin, and the attestation language gives a false claim something to hang on later. It does mean the system's credibility over the next year will be decided by what happens the first time a certified SMP-1 is credibly alleged to be an SMP-4 — whether there is a process, and whether anyone runs it.
For now the scale exists, one platform is using it, and a viewer has a number where previously they had a suspicion. A Screamify representative framed the platform's reasoning in plain terms: audiences are going to ask how these films were made, and it is better to answer the question on the product page than to be asked it afterward.
The rest of the industry has not responded. No other streaming service has adopted the system.
Comments (3)