On August 2, 2026, the European Union's AI Act reaches the deadline that converts a large slice of it from text into enforceable law across all 27 member states — and the portion switching on is the transparency chapter that lands squarely on the AI-companion and synthetic-intimacy businesses Afterglow tracks. From that date, providers must tell people when they are talking to a chatbot rather than a human, mark AI-generated images, audio and text, and clearly label deepfakes and machine-written articles published to inform the public. As Silicon Canals reported on July 18, 2026, the transparency obligations are "the cheap ones to meet" — while the tougher, rights-protective rules on hiring, biometrics and migration were quietly pushed back to December 2027.

Why It Matters

The EU is the world's most consequential rule-maker for consumer tech, and the "Brussels effect" means compliance features built for Europe frequently become global defaults. For AI-companion startups — many of them thinly capitalized and operating in a legal gray zone — a hard requirement to disclose bot status and watermark generated content is both an engineering cost and a potential conversion killer, since immersion is the product. For the broader sextech and adult-content industries, machine-readable provenance labeling on synthetic sexual media is exactly the infrastructure that anti-deepfake enforcement (and platform takedown systems) need to function, making August 2 a quiet but real inflection point for the entire synthetic-intimacy economy.

For the fast-growing AI-companion sector — Replika, Character.AI, Candy.AI and the wave of "girlfriend/boyfriend" apps — the disclosure mandate is the most directly relevant piece. Companies whose entire value proposition rests on an emotionally convincing, human-feeling conversation must now legally break the fourth wall for European users, surfacing that the "partner" on the other end is software. The same applies to AI-generated erotic and pornographic media: synthetic sexual imagery and audio distributed in the EU must carry machine-readable provenance markers, tightening the screws on the deepfake-porn pipeline that lawmakers on both sides of the Atlantic have spent the past year targeting.

The rollout is staggered and, as commentators noted, somewhat lopsided. The transparency rules are enforceable now, but enforcement machinery, harmonized standards, and the Commission's finalized guidance are still catching up, and the heavier obligations governing "high-risk" systems were delayed more than a year. Technology.org's July 17 breakdown emphasized that what "actually applies" on August 2 is narrower than headlines suggest — general-purpose AI transparency and labeling duties, not the full high-risk regime. Still, for any platform serving EU users, the labeling clock is now running.

This dovetails with parallel moves worldwide: China's AI-companion persona rules, Tennessee and Hawaii chatbot-safety laws, the UK's "intimate functionality" framing, and the U.S. DEFIANCE and TAKE IT DOWN acts. The regulatory center of gravity for synthetic intimacy is shifting from "should we regulate" to "how do we label and audit."

Sources

Update — 2026-07-20

Initial entry — story first created.