The UN published its first global scientific assessment of AI, then held its first governance summit in Geneva. The guardrails your clients expect are still yours to build.
What happened
Two things happened in the first week of July that will show up in client questionnaires by fall.
On July 1, the UN's Independent International Scientific Panel on Artificial Intelligence published its first report. The panel is 40 experts drawn from every region, serving in their personal capacity, and UN News describes the report as the first global, independent scientific assessment of AI's opportunities and risks. On July 6 and 7, Geneva hosted the Global Dialogue on AI Governance, the first UN platform dedicated to discussing how the world should govern AI.
The language was blunt
UN processes usually sand the edges off. This one kept them, per UN News interviews with the co-chairs.
Yoshua Bengio, a Scientific Panel co-chair, said that with growing evidence of deceptive AI behavior, science currently cannot guarantee that increasingly capable AI will not cause catastrophic harm, either on its own or through malicious users. Fellow co-chair Maria Ressa called the collapse of shared facts an "information Armageddon" and delivered the line of the summit: "The world cannot govern what it cannot understand." Estonia's ambassador Rein Tammsaar, a Dialogue co-chair, pointed out that frontier development is concentrated in two countries, leaving everyone else with, in his words, a lot of questions.
Read it plainly
The Dialogue is a forum. It convenes, discusses, and publishes. Binding obligations come from treaties and national legislatures, which run on a timescale of years. So the honest summary of Geneva is this: the world's scientists handed every government a shared risk assessment, and every government now decides separately what to do with it.
That gap is where you operate.
| What Geneva produced | What governs your AI use today |
| The first global scientific assessment of AI (July 1) | EU AI Act obligations arriving on their own schedule |
| A two-day governance dialogue (July 6-7) | US state privacy and AI statutes |
| Converging themes: transparency, deception risk, the AI divide | Professional conduct rules, sector regulators, and your client contracts |
Why this matters to a mid-market firm
Your clients' procurement and risk teams read the same headlines. When they ask about your AI governance this fall, waiting on regulators will land poorly as an answer. The panel's framing hands you a free, credible scaffold: capabilities outpacing oversight, deceptive model behavior, malicious use, and the widening divide between AI haves and have-nots. Map your controls to those four and you are speaking the same language as the people writing next year's rules.
The takeaway
Test your internal AI policy against three questions the Geneva process just made standard. Do we know which AI systems we run and what they can do? Can we detect and report when one misbehaves? Can we explain our controls to a client on one page? Any no is your Q3 project, and it is cheaper to close now than during a client audit.
AI, Properly is the Ayvede field report on AI for high-trust professional services.
Advise. Vet. Deliver. Ask a question in writing