EU AI Act — Full Enforcement Timeline (2024 → 2027)
The EU AI Act (Regulation 2024/1689) does not switch on as a single date. It applies in five phased windows. The milestone most AI video platforms search for — “EU AI Act implementation June 2026” — is actually August 2, 2026. This page lists every binding date and what each one obligates a platform to ship.
- August 1, 2024Entered into forceRegulation (EU) 2024/1689 published in the Official Journal
The AI Act formally entered into force. Most obligations are NOT immediately applicable — the regulation uses a phased application schedule keyed off this date.
- February 2, 20256 months inProhibited AI practices apply (Chapter II)
Article 5 prohibitions take effect: social scoring, real-time remote biometric identification in public for law enforcement (with narrow carve-outs), exploitation of vulnerabilities, and others. AI literacy obligations under Article 4 also start.
- August 2, 202512 months inGeneral-Purpose AI (GPAI) model obligations apply (Chapter V)
Providers of GPAI models become subject to transparency, technical documentation, and copyright-compliance obligations. Members of the AI Office and the AI Board are operational. Penalty rules in Articles 99 and 101 take effect.
- August 2, 202624 months in · TODAY+General application — Article 50 transparency obligations apply
This is the date most AI video platforms care about. Article 50 transparency obligations apply: AI-generated deepfakes depicting real or fictional persons must be machine-readably labelled as artificially generated or manipulated. Conformity assessment for high-risk systems listed in Annex III also kicks in.
- August 2, 202736 months inHigh-risk AI systems embedded in regulated products fully apply (Annex I)
High-risk obligations for AI systems intended to be a safety component of products covered by existing EU harmonisation legislation listed in Annex I (e.g. medical devices, machinery, toys) become fully applicable.
The June 2026 confusion
The widely-quoted Article 50 date is August 2, 2026 — exactly 24 months after entry into force on August 1, 2024. The “June 2026” phrasing that shows up in search queries appears to conflate implementation deadlines with national supervisory authority designation deadlines (most member states had to notify authorities by August 2, 2025; some published guidance during spring 2026). For the obligation itself, August 2, 2026 is the binding date.
What August 2, 2026 actually obligates
Article 50(4) requires that providers of AI systems generating synthetic image, audio, video or text content disclose the artificial origin in a machine-readable format. For content constituting a deepfake of an identifiable real person, the deployer additionally has to disclose the artificial nature of the content. Together these create a two-sided obligation: upstream provider labels the output; downstream deployer ensures the depicted person consented to the use.
Penalty exposure under Article 99 reaches €15 million or 3% of global turnover for transparency breaches, and €35 million or 7% for prohibited-practice or repeat violations.
Where ActorHub fits in this timeline
ActorHub is a pre-generation consent layer. Two integration touchpoints map directly to Article 50:
- Provider side — the SDK emits an ES256-signed trust signature on every generation that depicts a registered identity. The signature is the machine-readable evidence the article requires, verifiable at actorhub.ai/.well-known/jwks.json without ActorHub's cooperation.
- Deployer side — the consent check returns withdrawn whenever the depicted person has revoked consent on any integrated platform, so a downstream deployer cannot inadvertently push content past a withdrawal.
Integrate before August 2, 2026
Sandbox key in 24 hours. Production rollout in under two weeks for most platforms. We're approving the first 50 integrations manually — reply within 2 business days.