Hard Numbers: Stolen art, mathletes, DeepMind defection, Antitrust tussle, Labor shortages

​Midjourney artificial intelligence on a computer screen.

Midjourney artificial intelligence on a computer screen.

IMAGO/Action Pictures via Reuters Connect
16,000: A list of 16,000 names of artists that Midjourney allegedly trained its image generation model on has spread on the internet after leaking from an online court filing. The list purports to show that Midjourney trained its model on the works of artists like Frida Kahlo and Damien Hirst. The list is an exhibit from an ongoing class-action lawsuit against the AI company and was accidentally posted online before being removed.

25: A math-focused AI system is almost at the level of our species’ most advanced teenage mathletes. The model called AlphaGeometry reportedly scored a 25 out of 30 on the International Mathematical Olympiad, the premier high school math competition. The average gold medalist in the competition has scored a 25.9.

220 million: A pair of Google DeepMind engineers are reportedly ready to leave the company, in talks to raise $220 million to build a new large language model. The company, which would be based in France, is tentatively called Holistic.

2: The US government's effort to investigate Microsoft’s $13 billion investment in OpenAI is on hold. The problem: Two different agencies can’t decide on who should take the lead. This is the problem that’s arisen between the country’s two antitrust authorities, the Department of Justice and the Federal Trade Commission. Neither party seems ready to relinquish the job to the other.

11 million: Japan is estimated to face a labor shortage of 11 million people by 2040 as its aging population peaks and exits the workforce. While AI and robots may displace human labor in some economies, they might also help others — such as Japan — avert crises in industries such as construction and retail.

More from GZERO Media

Across North America and Europe, blackouts are becoming more common, often hitting when the demand peaks or when the weather turns extreme. The surging demand for power and new energy sources are putting pressure on the energy systems. Meeting today’s energy needs takes a flexible, pragmatic, “all-of-the-above” approach — drawing on all fuels and technologies. Learn how Enbridge is delivering reliable, affordable energy in uncertain times.

Amir Seaid Iravani premanent representative of the Islamic Republic of Iran speaks during the UN Security Council on June 24, 2025 in New York City.
John Lamparski via Reuters Connect

It’s not clear yet how much the US attack on Iran's nuclear sites this weekend set back the Islamic Republic's ability to develop atomic weapons, but experts say the airstrikes almost certainly threw a bomb into something larger: the global nuclear non-proliferation regime.

A pie graph showing the percentage of Americans in favor of having a third major political party.
Ico Oliveira

Remember when Elon Musk threatened to start his own political party during his spat with Donald Trump? It’s unclear how many Americans would switch their political affiliation to a Musk-run party specifically, but a plurality agree that they’d like another major political party to rival the Democrats and Republicans.

Open Call is the heart of Walmart’s $350 billion commitment to US manufacturing, supporting products made, grown or assembled in America. The pitch event represents a unique opportunity for selected entrepreneurs to meet face-to-face with Walmart merchants and earn a chance to get their products on store shelves nationwide. Last year, finalists from across the country represented 48 states, with entrepreneurs from over half these states receiving deals. It’s all a part of Walmart’s investment in American jobs and communities. Learn more about Walmart’s annual Open Call.

Last week, Microsoft released its 2025 Responsible AI Transparency Report, demonstrating the company’s sustained commitment to earning trust at a pace that matches AI innovation. The report outlines new developments in how we build and deploy AI systems responsibly, how we support our customers, and how we learn, evolve, and grow. It highlights our strengthened incident response processes, enhanced risk assessments and mitigations, and proactive regulatory alignment. It also covers new tools and practices we offer our customers to support their AI risk governance efforts, as well as how we work with stakeholders around the world to work towards governance approaches that build trust. You can read the report here.