The European Commission says the AI Act’s transparency obligations began applying on August 2, 2026. According to the Commission’s guidance, providers must tell users when they are directly interacting with an AI and must add machine-readable marks to detect AI-generated or manipulated content.
The same guidance says deployers must inform people when they are exposed to deepfakes, certain AI-generated public-interest content without human review, or emotion-recognition or biometric-categorization systems.
Reporting from TechTarget says many high-risk obligations, including employment and credit-scoring use cases, moved to December 2, 2027. The same reporting says some AI-generated-content labeling obligations remained on the August 2, 2026 timeline, with a narrower grace period for systems already on the market before that date.
The supplied sources do not identify any specific employer, product, or deployment as already noncompliant, and they do not state how many organizations have changed their internal processes. They also do not provide the full technical format of the “machine-readable marks” or the complete list of content covered by “certain AI-generated public-interest content without human review.”
What the Commission says now applies
The Commission’s guidance says providers must inform users when they are directly interacting with an AI and must add machine-readable marks so AI-generated or manipulated content can be detected.
The same guidance says deployers must inform people when they are exposed to deepfakes, certain AI-generated public-interest content without human review, or emotion-recognition or biometric-categorization systems.
What moved to a later date
TechTarget reported that many high-risk obligations, including employment and credit-scoring use cases, moved to December 2, 2027.
TechTarget also reported that some AI-generated-content labeling obligations remained on the August 2, 2026 timeline, with a narrower grace period for systems already on the market before that date.
What the supplied sources do not say
The supplied sources do not say which companies, agencies, or employers have already updated notices, labels, or user flows. They do not say that any specific hiring, applicant-communication, or employee-facing system has already been changed, and they do not identify any separate enforcement campaign tied to the August 2 start date.
The sources also do not state that product labeling or governance work is already required in every case, nor do they say that vendors have faced pressure beyond the reporting that many high-risk obligations were delayed to a later date.