Meta said on 28 July 2026 that it is signing the European Union's AI Act Code of Practice on Transparency of AI-Generated Content. The announcement came in a Meta Newsroom post by Markus Reinisch, the company's VP of Public Policy for Europe, setting out the reasoning behind a decision Meta had confirmed earlier that week. For anyone buying media or producing creative, the substance is narrower than the headline: Meta committed to a voluntary code and to interoperable detection work, and published no new advertiser-facing requirements.
What Meta committed to on 28 July
Meta's 28 July post states that the company is signing the EU AI Act Code of Practice on Transparency of AI-Generated Content, and exists mainly to explain why. Reinisch frames Meta's position around a single test: transparency measures should "provide clarity, rather than more confusion" for the people who encounter AI-generated content.
The post also carries a warning about how rules of this kind can misfire. Reinisch cautions against a situation where online content ends up carrying more and more separate labels and disclosures, which would overwhelm the people reading them while adding regulatory complexity for the providers building the systems. Meta closes by saying it will work with the AI Office and other partners "to keep the approach practical, interoperable and genuinely useful for the people it's designed to serve."
The detection and provenance work Meta points to
Meta dates its own work on this to February 2024, when it announced its approach to identifying and labelling AI-generated content on its platforms. Since then, according to the 28 July post, Meta has launched a research demo of a detection tool that helps people understand whether an image was made with Meta AI. Meta describes it as a research demo, and the post attaches no launch schedule or product timeline to it.
Meta also says it is working through Partnership on AI and the Coalition for Content Provenance and Authenticity, known as C2PA, to develop identification solutions that are "durable, technically-feasible and sustainable". The choice of venue matters more than it reads. A provenance signal is only useful if it survives a crop, a re-upload and a re-encode, and it only works at scale if every platform reads the same signal. Teams producing AI-assisted assets at volume for organic content marketing should plan for provenance metadata that travels with the file rather than staying inside the tool that made it.
What the 28 July post does not say
The gaps in Meta's announcement are worth naming, because most of the commentary around it will fill them in:
- It gives no enforcement date.
- It publishes no list of labels or disclosure formats.
- It sets no timeline for new advertiser requirements.
- It makes no reference to Thailand or to any market outside the EU context.
One piece of context sits outside the post. Under the EU AI Act's own timeline, the Act's transparency obligations become binding on 2 August 2026, and the Code of Practice is a voluntary instrument that signatories adopt rather than a law in itself. Meta's post does not state that date, or any date.
Timeline of what has actually happened
Three fixed points carry this story, and each one has a different source. The table separates them.
| Date | What happened | Who stated it |
|---|---|---|
| February 2024 | Meta announced its approach to identifying and labelling AI-generated content on its platforms | Meta's 28 July 2026 post |
| 28 July 2026 | Meta published its reasoning for signing the EU AI Act Code of Practice on Transparency of AI-Generated Content, having confirmed earlier that week that it would sign | Meta's 28 July 2026 post |
| 2 August 2026 | The AI Act's transparency obligations become binding; the Code of Practice itself remains voluntary | The EU AI Act timeline, not Meta's post, which states no date |
What this means for marketers in Thailand
Nothing in Meta's 28 July post requires a Thai advertiser to change a campaign. There is no new ad-account setting, no new disclosure field and no compliance deadline pointed at advertisers. A brand running Facebook and Instagram ads from Bangkok sits in the same position on 30 July as it did on 27 July.
What moves is the direction of travel. Signing a transparency code commits Meta to detection and provenance work it will apply inside its own systems, and platforms have generally settled on one global standard for labelling of this kind rather than maintaining a separate EU-only version. That convergence is an expectation, not something Meta stated in the post. If it holds, AI-assisted creative made for Thai audiences would carry the same provenance signals as creative made for Germany, without anyone in Bangkok opting into it.
The useful preparation is record-keeping, and it costs nothing to start now. A team that knows which assets were AI-generated, with which tool, and at which stage of production, can answer a disclosure question in an afternoon. A team without those records will be re-auditing a year of creative under a deadline set by someone else.
Questions marketers are asking
Does Meta's decision change anything in my ad account right now?
No. Meta's 28 July post published no new advertiser-facing requirements, no new account settings and no advertiser deadline. It is a policy commitment plus the reasoning behind it.
Is the Code of Practice mandatory?
No, the Code of Practice itself is voluntary. Under the EU AI Act timeline the Act's transparency obligations become binding on 2 August 2026, but Meta's post does not state that date or any enforcement date of its own.
Does this apply to campaigns I run in Thailand?
Meta's post does not mention Thailand. The Code of Practice is an EU instrument, and the post says nothing about how the commitment applies market by market. The reasonable expectation, and this is speculation rather than a Meta statement, is that a platform of Meta's size runs one global labelling standard rather than an EU-only one.
Will my AI-generated ad creative be labelled, and when?
Meta's post gives no list of labels and no timeline. What it does describe is a research demo of a detection tool for images made with Meta AI, plus work through Partnership on AI and C2PA on identification solutions Meta calls "durable, technically-feasible and sustainable".
The honest summary of 28 July is that Meta made a commitment and left the operational detail for later. That is a good window to get creative provenance in order while no deadline is attached to it. If you want a second read on how your Meta creative and organic content would hold up under a disclosure requirement, the time to look is before the requirement exists.
