A decade ago it would have been remarkable for the people who build frontier AI to sit beside heads of government; at the 52nd G7 summit, which opened in Évian-les-Bains on 15 June, it was the plan. Sam Altman, Dario Amodei, Demis Hassabis and Arthur Mensch are among roughly a dozen technology executives France invited into working sessions on AI regulation, digital infrastructure and the protection of minors online, according to a list released by the French presidential office.

The summit matters less for any single announcement than for what the guest list confirms: governments now treat advanced AI as strategic national capability, on a par with energy and semiconductors, and they have found a politically durable route into regulating it. That route is child safety, and its mechanics, age assurance and safety by design, run straight through the ad-tech plumbing every marketer depends on.

What has the G7 actually agreed on child safety?

G7 digital ministers agreed the substance on 29 May, five weeks before the leaders met. Meeting in Paris under France's presidency, ministers from Canada, France, Germany, Italy, Japan, the UK, the US and the EU set out their first common set of principles for protecting children online: digital literacy, safety built in by design, curbs on recommender systems that drive excessive engagement, and the load-bearing requirement, robust and privacy-preserving age assurance. The European Commission has said the principles draw on the Digital Services Act and the AI Act.

Anne Le Henanff, France's digital minister, was blunt about the intent, saying the declaration would let regulators be "far more demanding" of platforms and "leave them no choice but to change their way of working". That ministerial text serves as a preparatory document for Évian, where child protection sits on the leaders' agenda alongside digital infrastructure. The leaders' declaration was still developing as the summit opened, and the earlier ministerial meeting had reached only a limited deal, with divisions over AI's energy demand left unresolved.

Why does a child-safety rulebook reach B2B marketing?

B2B audiences are adults, which makes a child-safety framework look irrelevant to a SaaS marketing team, and that reading is the trap. Age assurance and data-sharing obligations are being built into platforms by design, which changes the plumbing for everyone: more sign-up friction, more privacy-preserving defaults, and fewer of the behavioural signals that power lookalike audiences and retargeting. Related proposals go further, treating profiling-based recommendations to minors as editorial activity and restricting targeted advertising to them. As that logic generalises, behavioural targeting gets harder across the board, and first-party data and intent-led content get correspondingly more valuable.

The same principles press for provenance and authentication of AI-generated media to curb abuse and non-consensual imagery. Once that is mandated for the worst cases, disclosure and "made with AI" labelling tend to harden into the default for all synthetic content. For marketing teams running AI-generated imagery, voices or synthetic video, provenance is shifting from a quality question to a compliance one, worth getting ahead of as a credibility signal rather than treating as a burden.

What should marketers take from AI's seat at the top table?

Marketing teams can read the summit as quiet validation rather than a fresh threat. When G7 leaders treat AI adoption as a growth priority, the lingering question of whether a team should be using these tools at all loses its force; the same 29 May package bundled an SME readiness tool, built with the OECD, to help smaller firms adopt AI faster. The conversation moves from permission to proficiency.

That framing is worth holding against the direction of travel. The G7 launched its Hiroshima AI Process in 2023 to build an international framework for trustworthy AI, and followed it with an "AI for prosperity" statement in 2025; the language has steadily shifted from safety and guardrails towards adoption and growth. The Évian agenda continues that drift, which is the quieter story behind the child-safety headlines: regulators are going firm on the politically unanswerable case, minors, while easing off the harder safety language elsewhere just as AI adoption accelerates.

The regulatory direction also rewards a strategy marketers can control. As third-party signals thin out, first-party data, owned audiences and content credible enough to be cited by both search engines and AI assistants become the more durable investments. One note of realism sits underneath the optics: ministers acknowledged AI carries "certain risks" and may be misused, but committed to an "innovation-friendly approach" to addressing them. There is no single rulebook coming. Europe regulates from principle, the US resists tighter rules, and the result is a patchwork marketing teams operating across markets will have to navigate one jurisdiction at a time.

The Évian summit runs until 17 June. Whether the leaders' declaration hardens the ministers' principles into firm commitments or leaves them as direction of travel will set the pace at which any of this reaches a marketing plan, but the platforms have already been told to change how they work.