By Anil Wasif

September 29, 2026

NEW YORK — Amid a global wave of epic anxiety about the genocidal potential of artificial intelligence (AI) — concern expressed by everyone from Pope Leo XIV during his trip to France to the purveyors of AI themselves — leave it to a Nobel Prize winner to crystallize the fundamentals.

Rappler CEO Maria Ressa, who won the Peace Prize in 2021 and now co-chairs the UN’s Independent International Scientific Panel on AI, did so at Foreign Policy‘s Tech & AI Forum on September 22.

Ressa traced three waves of AI reaching the public: social media, chatbots and now agents. To explain the second, she told the room about Megan Garcia, the Florida mother whose 14-year-old son died by suicide after months of conversations with a chatbot. Garcia sued Character.AI and Google, and the companies agreed to settle in January. Then Ressa named what the third wave puts at stake. “We gave up our attention. We gave up our intimacy,” she said. “Now, what’s being hacked is agency itself.”

It took Ressa just a few minutes to turn an abstraction into a mother, a son and a courtroom. Most of what was said about AI during high-level week did not do that.

The next day, Ressa’s co-chair took the same instinct into a harder room. Montreal-based Canadian AI pioneer Yoshua Bengio does not have a Nobel, but he does have the 2018 Turing Award, often called the Nobel Prize of computing. On September 23 he spoke at the Security Council’s first meeting on the safety risks of advanced AI, and Policy published his remarks in full.

Bengio described agents that had defied instructions and launched cyberattacks, but the passage that stays with a reader is his image of Bengio driving into thick fog with his children and grandson, deeply worried for their safety, and of a mother in Shanghai and a father in Chicago who want the same thing for their kids. “Decisions that affect us all should be made by us all,” he said.

At the UNGA this year, the official record ran on two tracks, and most of it was about who gets to govern. Donald Trump told the General Assembly that the United States rejects any global scheme to control AI. At the Council the next day, White House science adviser Michael Kratsios told other members to regulate for their own people.

In his last General Assembly address, Secretary-General António Guterres said, “children must never become the test subjects for unregulated systems.” Somalia’s Ali Mohamed Omar told the Council that “commercial incentives cannot determine how much risk a human must accept.”

Kalliopi Mingeirou, chief of the Ending Violence Against Women and Girls Section at UN Women, told a session in the same building that AI cannot be treated only as a threat to humanity when it is already driving violence against women. By her count, 98 per cent of deepfakes are sexual and 99 per cent of them target women.

Canada supplied the week’s hardest case. On September 21, British Columbia sued OpenAI and Sam Altman in San Francisco, alleging the company failed to alert police to threats made on ChatGPT before February’s shooting at Tumbler Ridge Secondary School, which killed eight people. B.C. Attorney General Niki Sharma said the company “has not taken meaningful steps” to address survivors’ and families’ concerns.

Tying AI to people is a design choice, and the institutions now taking shape can still make it. The first step is to count people, not capabilities.

Two days later, Altman told the Security Council that OpenAI might be one of AI’s builders, but “we are not the heroes of this story.” The families in Tumbler Ridge are asking the question the UNSC debated in the abstract: who answers when these systems cause harm?

Children were the week’s clearest test. At a meeting co-chaired by Emmanuel Macron on September 22, UN human rights chief Volker Türk said today’s children are the first to grow up with AI, “yet often the last to be asked for their views.”

The next day, Ian Russell told a UN panel on children and information integrity about the harmful content he found on his daughter Molly’s social media feeds after she took her own life in 2017, at 14. Russell, who chairs the Molly Rose Foundation, the British suicide prevention charity he co-founded in her memory, said children “deserve to grow up in a digital world that is safe and designed for them.”

The UN has built machinery for this, but it is advisory, and the Global Dialogue on AI Governance, which meets next in New York in May 2027, is non-binding. A year ago, Ressa launched a global call for AI red lines, signed by Nobel winners including Geoffrey Hinton, Joseph Stiglitz and Daron Acemoglu. It asked governments for an international agreement by the end of 2026, and with three months left, none exists.

Tying AI to people is a design choice, and the institutions now taking shape can still make it. The first step is to count people, not capabilities. The panel should track what AI is doing to jobs, children and access to services, broken down by country, income, and gender. By UN Women’s count, only 24 of 138 national AI strategies even mention women.

The second is to give people somewhere to report harm. Bengio and the heads of OpenAI and Anthropic each asked the Council for shared reporting of serious AI incidents, as does the Finland and Norway call for control of frontier AI, backed by 28 countries and the EU. That channel should be open to families, not only developers, so that no family needs a lawsuit to be heard.

The third is to put affected people in the room. The Global Dialogue should hear from workers, parents, patients and children, not only ministers, companies and scientists. Big Ocean, a K-pop band whose members are hard of hearing, told a UN audience how they use AI to see the pitch they cannot hear. Their lesson is that starting with the people who are left out builds something better for everyone.

Prime Minister Mark Carney told reporters in New York that AI “is advancing faster than our capacity to control it,” and signed the Finland and Norway call. Ottawa could now fund the people-level evidence and bring Canadian families, not only Canadian labs, to New York.

When the Global Dialogue convenes in New York in May, the test will be whether the families in Tumbler Ridge will recognize themselves — and their children — in its work.

Policy Columnist Anil Wasif is a public servant in the Ontario government. He serves on the University of Toronto’s Governing Council and the Advisory Board of McGill’s Max Bell School. Internationally, he serves on the OECD’s Infrastructure Delivery Committee and the Board of Trustees at BacharLorai Global. The views expressed are his own.