AI policy has become a three-way argument. Governments want control, companies want room to build, and the technology keeps moving before either side finishes its sentence.
Greece’s Prime Minister Kyriakos Mitsotakis arrived in San Francisco on Monday night to promote Greece as a destination for technology companies and talent. He also admitted that he does not have answers to many of the questions leaders around the world are debating, describing his Bay Area trip as partly a fact-finding mission.
He spent the morning touring Tesla and Sequoia Capital, among other stops, before planning to travel to the U.N. General Assembly in New York this week. At an event hosted by Endeavor Greece, Mitsotakis was pitching anyone in tech who might consider setting up shop in Greece — a familiar political ritual, except this time the sales pitch comes with a national AI anxiety problem.
Greece Wants the Builders and Fears the Consequences
Greece is set to regain developed market status next year from MSCI, while its debt is being paid down at a record pace. Its 10-year bond yield sits around 4.3%, below the roughly 5% yield on U.S. Treasuries, after Greek yields topped 40% during the 2012 debt crisis.
That turnaround supports Mitsotakis’s technology pitch. Much of the roughly €36 billion Greece received from the EU’s post-COVID recovery fund went into digital infrastructure, including an online portal for government paperwork and a new supercomputer assembled with Hewlett Packard Enterprise in Lavrio.
Greece has changed how stock options are taxed, loosened labor laws, and offered returning Greeks lower taxes for up to seven years. It also has a growing array of visa programs, though processing speed remains an issue, while public universities have seen a stark turnaround and now spin out startups.
Mitsotakis wants to win back talent lost during Greece’s debt crisis. He suggested that founders consider hiring in Greece as American work visas become harder to obtain. The country is building the welcome mat while still trying to stop the guests from wandering into the nursery.
Greece will institute a ban on social media for children under 15 in January. Mitsotakis also raised concerns about AI chatbots’ addictive nature and their effects on society, asking, “What does it mean for our kids to grow up with digital companions or boyfriends or girlfriends?”
Greece has run an education pilot with OpenAI aimed at reducing teachers’ administrative work. Mitsotakis sees promise in personalized AI tutors but warned that AI should not replace basic skills learning, even as students already use chatbots to do homework. “Sometimes I feel that we’re already fighting yesterday’s battle,” he said.
Governments Want Rules While Companies Write Their Own
Andy Burnham called for a universal set of rules to monitor AI growth and said the UK should lead in developing global principles and standards for frontier AI. He said the UK would use its G20 presidency to push for international agreement on AI safety, transparency, and standards, arguing that decisions about the technology must not be left to chance.
“Fundamental decisions on this technology are too important to be left to chance. They must be taken by elected governments who serve working people,” Burnham said. He also acknowledged that AI could disrupt jobs, while pointing to its potential in healthcare, economic growth, education, and defense.
Burnham’s speech at the UN came hours after Donald Trump said the United States would resist attempts to rein in super intelligence. Trump said the US would encourage super intelligence, watch it through the Department of Justice, and continue leading the world in the field “safely and responsibly.”
Burnham said the UK is deeply tied to the US, EU, China, India, and Brazil, describing the country as able to bring people together in a divided world. Critics, including former Cabinet minister Darren Jones, said the UK should have joined a Norwegian and Finnish initiative for global AI rules, which has support from more than 20 countries.
Liberal Democrat leader Sir Ed Davey called for a global treaty to pause super-intelligent AI and establish an international AI safety agency. The UK has instead told AI firms that full visibility over new models is expected, with legislation threatened if necessary.
That gap between public oversight and private control is also widening in the United States. The US White House has not pursued a federally run AI regulation organization, Congress members show little interest in legislation, and China’s Ministry of Foreign Affairs accused Americans of “fear mongering” over AI regulation. A China state-run newspaper called Dario Amodei’s rhetoric a “Cold War playbook.”
Amodei, Anthropic’s CEO, wrote a nearly 4,000-word essay calling for slower AI development and international collaboration on safe deployment. His plan would coordinate companies and governments, and rival executives Sam Altman and Elon Musk have endorsed it; it also argues that pacing the AI frontier could help the US widen its lead over China within 3–5 years.
Not everyone wants governments near the controls. Meta CEO Mark Zuckerberg delayed the release of its AI model Muse for safety and security reasons, while implying that market pressure can push companies toward safer models without government intervention. Reddit co-founder Alexis Ohanian criticized the industry as tone deaf about AI risks but said companies may still manage safety independently.
Google DeepMind co-founder Shane Legg offered the clearest warning: “We’re living in a period now where capabilities are advancing very, very quickly. But we can’t let capabilities get ahead of safety.” OpenAI, Anthropic, and other major AI companies are creating a private “self-regulatory body,” while Cohere CEO Aidan Gomez accused American AI companies of forming a “cartel” to control safety rules.
The UK House of Commons Business, Innovation, Science and Trade Committee has invited Meta, Google, OpenAI, and Anthropic to give evidence on AI safety testing before deployment. The hearing is scheduled for October 13, 2026, with companies asked to confirm attendance and provide written responses by September 29.
High-level executives are debating whether incidents such as the Hugging Face incident suggest AI is nearing chaos. The argument has moved beyond whether guardrails are needed. As Gomez put it, “AI needs guardrails. That is not the dispute and never has been. The dispute is over who writes them, who gets to participate and whose interests the rules are protecting.”
Based on
- ‘We’re already fighting yesterday’s battle’: Greece’s prime minister gets candid about AI — techcrunch.com
- Burnham says AI safety should be an issue for governments and not left to chance | The Independent — independent.co.uk
- Is the AI safety debate about safety or control? | TechCrunch — techcrunch.com
