Meta signs EU AI transparency code as August compliance deadline looms

Meta has confirmed it will sign the EU AI Act Code of Practice on Transparency of AI-Generated Content, joining Google and other major platforms in committing to the voluntary framework ahead of mandatory enforcement rules taking effect on 2 August 2026.

The transparency obligations require providers and deployers of AI systems to mark, label, and watermark AI-generated content, with non-compliance carrying fines of up to €15 million or 3 per cent of worldwide annual turnover.

The European Commission encouraged companies to submit signatory forms by 18:00 CET on 22 July 2026 to appear on the initial list of signatories, though organisations can sign at any time, including after 27 July 2026.

Why Meta is signing now

The move builds on Meta's approach to identifying and labelling AI-generated content announced in April 2024, when the company began adding "AI info" labels to video, audio, and image content when it detected industry-standard AI image indicators or when people disclosed they were uploading AI-generated content.

Starting in May 2024, AI-generated user content featured a "Made with AI" label, with detection coming from industry-shared signals and user self-disclosure. The company later rolled out changes so labels better reflected the extent of AI used in content, renaming them "AI info" labels.

Meta has also launched a research demo of a detection tool to check whether an image carries a Content Seal watermark, an invisible watermarking system that remains in place even when cropped, compressed, resized, or screenshotted.

Signing the Code of Practice offers streamlined compliance, greater predictability, legal certainty across the EU, and reduced administrative burden, with future enforcement focused on monitoring adherence to the code. Companies that decline to sign will need to demonstrate compliance through other means and may face greater scrutiny.

What the code actually demands

The EU AI Act requires providers to implement machine-readable marking on AI-generated audio, image, video, and text. But the Code acknowledges that under the current state of the art, there is no single solution that can provide sufficient reliability.

In most cases, signatories will need to implement a multi-layered marking approach including at least two methods: metadata, where the content format supports it, providers must record AI-generation status in the metadata, digitally signed and time-stamped in a tamper-evident manner, and imperceptible watermarking such as Google SynthID, deployed simultaneously, as neither technique alone meets the statutory requirements.

Section 1 of the Code sets out provenance requirements for providers offering generative AI systems, while Section 2 sets out transparency rules for deployers of AI tools, such as requirements to visibly label deepfakes and AI-generated text on matters of public interest.

Code of Practice signatories commit to visually disclose the existence of AI-generated markings using a set of generalised icons provided by the EU AI Office, along with information on whether the content was created or manipulated using AI.

The wider context: platforms, standards, and enforcement

Google confirmed it is signing the Code, building on its 2025 signing of the General Purpose AI Code of Practice and aligning with its work to adopt and accelerate the C2PA industry standard and to develop AI transparency tools including SynthID.

The Content Credentials specification is hosted by the Coalition for Content Provenance and Authenticity (C2PA), which has grown to a collaboration with hundreds of companies led by Microsoft, Adobe, Intel, BBC, Truepic, Sony, Publicis Groupe, OpenAI, Google, Meta, and Amazon. C2PA joined the efforts of The Content Authenticity Initiative and Project Origin to focus exclusively on the development of open, global technical standards for establishing content provenance and authenticity.

Unlike most other provisions of the AI Act, the transparency obligations are instantly applicable to all AI systems within scope, regardless of when the system was placed on the market or put into service, meaning existing AI systems must comply from day one.

Pre-existing generative AI systems have until 2 December 2026 to comply with this specific obligation.

What this means for social media marketers

The signing puts Meta's platforms in line with evolving regulatory expectations across major markets. The EU AI Act requires businesses to clearly label content that has been significantly generated by AI starting 2 August 2026, and California's CCPA and the EU AI Act require AI-generated files to contain machine-readable metadata.

Platform requirements are diverging. Meta requires disclosure for synthetic media on political and social topics; Google requires an 'AI Generated' label on all synthetic content; TikTok uses automated detection systems that prompt creators to label AI-generated content at upload; YouTube requires creator-side disclosure.

For marketers working across borders and platforms, the practical takeaway is straightforward: Meta, TikTok, and YouTube all auto-detect AI through C2PA Content Credentials, so many tools label creative automatically with no action needed. But for ads about social issues, elections, or politics, Meta requires advertisers to disclose when the ad contains photorealistic synthetic media, including to depict a real person saying or doing something they did not, or a realistic-looking person or event that does not exist.

Early data shows AI labelling drags down engagement on the metrics that drive paid social ROI, with a 2026 study finding that labelling content as AI-generated or AI-enhanced reduced both affective and behavioural engagement compared to human-created content, with the sharpest drop on emotional creative. That performance gap makes compliance a strategic issue, not just a legal one.

The code's multi-layered marking requirements and the August enforcement date mean social media teams need to audit how content is created, ensure tools embed the required metadata and watermarking, and confirm that disclosure processes align with both platform policies and EU law. Signing the code does not eliminate the obligation. It simply provides a clearer compliance pathway before enforcement begins.