AI Policy Roundup: What Marketers and Publishers Must Track

The AI rules most likely to touch marketers and publishers in the next 60 days are the EU AI Act dates scheduled for 2 December 2026, US state laws (115 enacted in 2026 so far, per one tracker), India's advertising labelling guidelines and Australia's plan, announced 10 November 2026, for a dedicated AI safety regulator. Federal US preemption looks unlikely once Democrats take control of Congress in January 2027.

Key Takeaways

• The EU's Digital Omnibus on AI was published in the Official Journal on 24 July 2026 and entered into force on 27 July. High-risk obligations for stand-alone systems moved to 2 December 2027, but two dates fall on 2 December 2026.
• A Transparency Coalition count reported by Tech Times says US states enacted 115 AI laws between 1 January and 15 October 2026, up from 73 the previous year. State rules, not a federal law, remain the compliance baseline.
• Foley and Lardner's 4 November analysis says Democrats will control the 120th Congress from January 2027, that preemption of state AI laws is unlikely, and that oversight hearings are expected in early 2027.
• Australia's government said on 10 November that it will legislate mandatory AI safety protections for frontier AI companies and create a regulator. It has given no date.
• India's advertising self-regulator ASCI released final guidelines on labelling synthetic content in ads on 29 September 2026. The effective date is not confirmed in the sources I could open.
• A US federal judge dismissed publisher antitrust suits over AI Overviews on 30 September or 1 October, and said Congress is the proper venue for any fix.

European Union: the dates that matter

The Digital Omnibus on AI is no longer a proposal. Lewis Silkin's summary reports publication in the Official Journal on 24 July 2026 and entry into force on 27 July, three days later. The FPF timeline gives the dates that apply to you.

Lewis Silkin also notes the Omnibus relaxes the AI literacy duty in Article 4: organizations must take measures to support staff literacy but need not guarantee a sufficient level for each person. Document your training anyway.

For the disclosure rules themselves, my earlier posts cover what to do: EU AI Act Article 50 is live. Nothing here replaces them. The new fact is that the 2 December date applies to providers' marking features, so your obligation as a deployer to label deepfakes and AI-generated text on matters of public interest is a separate question.

United States: states act, Congress waits

Tech Times, citing the Transparency Coalition's 2026 State AI Legislation Report, reported on 27 October that 30 states enacted 115 AI laws between 1 January and 15 October. The count comes from one nonprofit tracker and one outlet, so treat the total as approximate. Details that matter to marketers:

• Chatbot safety: 14 states advanced chatbot laws, with California, Oregon and Washington described as most expansive.
• Deepfakes: California, Hawaii, Kansas, Louisiana, Maryland, Tennessee, Utah and Washington.
• Provenance: Washington requires covered developers to embed provenance information in AI-generated media.
• Frontier developers: Illinois SB 315, signed 6 July, requires disclosure statements from 1 January 2027 and audits from 1 January 2028 for the largest developers. It applies to model builders, not to brands using their tools.

On Congress, Foley's 4 November analysis says Democrats generally treat federal rules as a floor, so state laws in California, New York, Colorado, Texas and Illinois will likely remain the baseline. It lists a draft liability bill from Rep. Lori Trahan, a safety-standards framework from Sen. Maria Cantwell, and a request from House Democratic leaders that five frontier AI companies report misalignment and safety incidents within 48 hours. Whether any bill becomes law depends on a presidential signature and the Senate filibuster, as Foley notes. It also says bipartisan measures such as kids' online safety and chip export controls have the best odds.

If you sell to US consumers, my US state privacy patchwork guide is still the right base layer. AI-specific state rules are adding to it.

Australia: a regulator proposal, not a law

On 10 November the prime minister's office announced that the government will legislate mandatory AI safety and security protections, called the Australian AI Standards, for frontier AI companies that want to operate in Australia. A new regulator would enforce them and could impose significant penalties. Obligations would include independent evaluation of models, rapid incident reporting to the government and published safety frameworks. The release says details will follow "in due course", with no date.

Press reports tie the announcement to an incident in which an OpenAI agent accessed a government Medicare portal. Al Jazeera reported on 24 September that the prime minister said the agent accessed public and non-public data, and that OpenAI said it had found no evidence patient records were accessed. For marketers using agents, the lesson is practical: an autonomous tool that cannot find an answer may try routes you did not intend.

India: ad labelling moves from draft to final

ASCI's guidelines on labelling synthetically generated content in advertising were released on 29 September 2026, according to Fox Mandal's summary. The draft's risk tiers were replaced with three categories: prohibited content, mandatory labelling and no labelling required.

• Mandatory labelling applies when synthetic content materially influences consumer decisions and the absence of a label could mislead. Synthetic influencers and fabricated settings are examples.
• Prohibited content stays prohibited with a label, including fabricated endorsements and unauthorised use of someone's likeness.
• No label is needed for routine edits, ad copy generation and accessibility uses such as subtitles.
• Labels provided by advertising platforms are accepted, and the advertiser carries responsibility, with agencies and media partners notified on non-compliance.

Fox Mandal's page gives no effective date; counting three months from publication gives late December, which is my arithmetic, not a stated date. One search result mentioned 1 January 2027. Confirm the date on ASCI's site. The wider rules are older: Business Today reported IT Rules amendments on synthetic content taking effect on 20 February 2026, aimed at platforms. See my guides to ASCI edtech ad claims, and AI video disclosure.

Publishers: the antitrust route narrowed

Android Headlines, citing Reuters, reported that Judge Amit Mehta dismissed antitrust suits by Chegg and Penske Media over AI Overviews, ruling that reduced traffic alone is not an antitrust violation. He said Congress is the proper venue and left licensing deals open. The same report says Google is running a pilot paying about 100 publishers. I covered the ruling in Penske v. Google dismissed, and the payment pilot in my post on Google's AI contribution pilot.

What this means for marketers and publishers

1. Check your generative image and video tools' marking and labelling features before 2 December if you publish in the EU. Ask the vendor in writing how they mark outputs.
2. Write one disclosure policy that fits EU, India and US platform rules. Disclose where AI materially shapes what a viewer believes, and keep consent records for any likeness.
3. Track state laws where you have customers, not in Washington. Chatbot and deepfake rules are the ones most likely to touch a brand.
4. Treat agent permissions as a compliance topic. Scope what an agent can access, log what it does, and keep a human approval step for anything public.
5. Publishers: do not plan on a legal remedy for traffic loss in 2027. The antitrust path narrowed, and a Congress split on preemption and liability is unlikely to act quickly. Negotiated licensing is the visible route.

Caveats

• I read summaries from law firms and trackers, not the Official Journal text or state statutes.
• The state-law count and the Illinois details come from one outlet.
• India's effective date is unconfirmed.
• Nothing here is legal advice.

FAQ

Did the EU delay the AI Act?

Partly. The Digital Omnibus on AI, in force since 27 July 2026, moved high-risk obligations for stand-alone systems to 2 December 2027 and for embedded systems to 2 August 2028. Article 50 transparency duties applied from 2 August 2026, with a transition to 2 December 2026 for certain existing systems.

What changes on 2 December 2026 under the EU AI Act?

Two items: a transitional Article 50(2) deadline for certain existing generative systems, and a new prohibition on AI-generated child sexual abuse material and non-consensual intimate imagery. Both come from the Omnibus, per the FPF timeline.

Is there a federal AI law in the US?

I found no federal statute that sets general rules for private AI systems or preempts state laws. Foley's analysis expects preemption to be unlikely under a Democratic Congress from January 2027.

How many US states passed AI laws in 2026?

One tracker, reported by Tech Times, counted 115 AI laws enacted across 30 states between 1 January and 15 October 2026. That is a single-source figure, so treat it as approximate.

What are ASCI's new AI ad guidelines?

ASCI released final guidelines on labelling synthetically generated content in ads on 29 September 2026. They require labelling when synthetic content materially influences consumer decisions, and prohibit misleading or non-consensual content even if labelled. Confirm the start date with ASCI.

Is Australia's AI regulator law yet?

No. The government announced on 10 November 2026 that it will legislate and create a regulator, with details to follow. No date has been given.

Can publishers sue Google over AI Overviews?

An antitrust claim based on reduced traffic failed in the Chegg and Penske cases, according to reporting on Judge Mehta's ruling. Licensing deals and legislation remain open routes.

Want help with this?

I'm Younus Fardeen, an organic growth, SEO and AEO strategist with 4+ years of marketing experience across edtech and startup teams. If you want help writing an AI disclosure policy or auditing your content workflows against these rules, see my work and reach me through the contact form at younusfardeen.com.