What it means for us

A first-time founder uses AI to build a mental-health platform for foster children

Michelle Turner is a foster parent who had never started a company. She used AI tools to teach herself the startup world, write a business plan, and sharpen a venture-capital pitch — not to replace her knowledge of the problem, but to translate it into language that rooms full of investors could act on. Her company, Here Now Health, launched in January 2025, now employs sixteen people, and provides Medicaid-funded mental-health counselling for children entering the foster system across three states. In a world where we all need to be communicators, this shows how AI can help us get past some of the usual barriers.

Geneva becomes the stage for AI governance issues

The UN’s first Global Dialogue on AI Governance opened here in Geneva on Sunday, with Secretary-General António Guterres warning that AI is developing faster than the rules designed to govern it. The Dialogue runs July 6–7, immediately alongside WSIS Forum 2026 and the ITU AI for Good Global Summit, which runs through July 10. One line that stood out to me in the speech: “When a child is harmed, the answer must never be ‘the algorithm did it.’”

The UN scientific panel says AI could widen inequality if governance lags

A panel of forty independent experts released a preliminary global AI assessment ahead of the Geneva Dialogue. They outlined the benefits… and the risks — children’s safety, misinformation, autonomous systems, and the concentration of AI power in a small number of organizations and countries. The Guardian’s framing: AI development may benefit those with infrastructure, data, and governance capacity, and leave everyone else with the bill.

AI for Good Global Commission launches

The UN and ITU have launched an AI for Good Global Commission, co-chaired by Salesforce CEO Marc Benioff and Rwandan President Paul Kagame, with a first meeting set for July 8 in Geneva. The mandate covers AI infrastructure, health, education, food security, disaster response, and trust. Plenty of upside, but Axios gets to the crux: “World governments are miles apart on how AI should be regulated, even as many countries agree that democratic values should govern the technology.”

Note: I’ve done some work with ITU in the past. Nevertheless, a story to follow due to its implications on AI overall.

Zuckerberg tells staff AI agents are coming along slower than the hype

Meta’s Mark Zuckerberg told employees that AI agent technology is developing more slowly than expected. Reuters reported that Meta had restructured teams, cut jobs, and moved people into AI roles — and the payoff hasn’t arrived yet. Zuckerberg still expects progress in three to six months. So even Meta, with a budget that makes most national GDPs look like a round of Heinekens, has discovered that “agentic transformation” doesn’t become real just because you put it in your bestest dream journal. The lesson here is mundane but clear: you can likely get away with “we’re working on it,” but pretending it’s already working is, well, complicated.

AI is changing how your organization gets described

Business Insider’s Cannes roundtable made the case that AI systems are now synthesizing your organization’s reputation from earned media, reviews, third-party content, structured data, and trusted sources — not just your “About Us” page. Digiday: AI search isn’t just an SEO issue, it’s a brand coherence issue. Your approved language is one witness in a crowded courtroom where you don’t control the other witnesses. The practical question for communications teams is not “are we visible?” It’s “are we being accurately described by systems we didn’t write, in contexts we can’t see, to audiences who trust the answer?”

Google draws a line: gaming AI answers is spam

Google updated its spam policies to cover attempts to manipulate AI Overviews and AI Mode — the generative search features now prominent in most results. Search Engine Land confirmed it: tactics designed to game AI-generated answers can violate spam policy. The Verge described the worst cases as “recommendation poisoning.” There is a legitimate version of boosting AI visibility: clear positioning, credible third-party validation, structured content, expert commentary, consistent proof. Then there is the other B.S.: content engineered to trick AI into recommending you.

Consumers are growing tired of AI-generated content — and they can smell it

Kantar research found that more than half of consumers surveyed believe AI ads increase the risk of misleading or fake content, and more than 40% said AI-generated advertising makes them uncomfortable or irritated. Separately, only 27% of influencer-led ads in the study clearly connected back to the brand. Clearly, audiences are developing a radar. A lesson for communicators? Generic content, disconnected from genuine brand voice, will underperform regardless of how efficiently it was produced.