ETH AI Policy Summit Opens Under Council of Europe Auspices
ETH Zurich’s Center for Law & Economics and RegHorizon open the annual AI Policy Summit on September 29–30, 2026 in Zurich, under the auspices of Council of Europe Secretary General Alain Berset, to convene policymakers, business, civil society and academia on capturing AI benefits while limiting risks. The 2026 themes—policy for innovation and competitiveness, policy for transition and transformation, and the road to Geneva 2027—sit inside the wider Zurich AI Festival backed by the ETH AI Center, Greater Zurich Area, city and canton authorities.
Filed under Policy and dated September 29, 2026, this AI4Switzerland briefing treats the summit as Swiss AI-governance convening news distinct from yesterday’s Apertus–Proton Lumo integration. Switzerland still lacks an omnibus AI statute and is drafting a consultation text toward early 2027 to implement the Council of Europe AI Convention; a Zurich-hosted dialogue with Berset’s patronage therefore matters for how cantons, banks and model hosts such as Swisscom or Infomaniak translate open models like Apertus into accountable practice before Geneva’s next multilateral markers.
Why it matters: Swiss firms already ship sovereign stacks while federal law trails. Multi-stakeholder summits can surface workable rules—but only if transparency, procurement clauses and human oversight survive the communiqué stage.
What it means in practice
Swiss CIO and public-affairs leads should inventory which AI programmes need Council-of-Europe-aligned documentation; demand named owners for Geneva 2027 preparatory notes; assign an owner for comparing summit takeaways to the early-2027 consultation draft; run time-boxed reviews of Apertus and other Swiss-hosted assistants against emerging norms; and prefer designs that keep humans on high-risk decisions. Anchor the summit to Switzerland’s early-2027 AI consultation plan and Apertus 1.5 inside Proton Lumo.
Caveats come first. A two-day summit is not legislation; festival branding can outrun statutory detail; and Council of Europe auspices are not EU AI Act transposition. AI4Switzerland therefore presents the gathering as directional governance context until published session outcomes appear.
What to watch next: whether summit briefs name concrete Swiss drafting options; how banks and cantons react; and links drawn to Alps-hosted open models. Readers can continue on the AI4Switzerland homepage, or browse the Newsroom for additional briefings.
Bottom line: treat this update as orientation, not instruction. Swiss AI policy is being argued in Zurich meeting rooms ahead of 2027 drafts and remains early. Organizations that benefit most will map summit themes to contracts now, keep humans on risk gates, and refuse to confuse a festival flagship with finished law.