World leaders and technology moguls convened in Seoul this week for the latest international AI safety summit, achieving the rare feat of agreeing that artificial intelligence should not be allowed to accidentally obliterate humanity. A bold stance, considering many of the companies involved are also racing to create machines that increasingly resemble omniscient digital sorcerers.
The two-day event, co-hosted by the United Kingdom and South Korea, drew notables including UK Prime Minister Rishi Sunak, European Commission President Ursula von der Leyen and officials from the United States and China. That’s right, the geopolitical equivalent of a family dinner where even the feuding cousins show up and attempt not to knock over the table.
The summit concluded with a non-binding agreement from sixteen countries and the European Union, pledging to develop safety frameworks for advanced AI models. A non-binding agreement, for those unfamiliar with diplomatic code, is a bit like raising one’s eyebrows and saying “we really should do something about this,” then calling it a job well done.
Perhaps most eyebrow-raising was Tech Bro Exhibit A, Elon Musk, who took to the stage in a virtual fireside chat to warn of the “scary” power of AI and urge caution. This from the same man whose companies are knee-deep in developing AI-driven cars and humanoid robots, suggesting that the fear he feels may exist in the curious zone between existential dread and marketing strategy.
Meanwhile, Google DeepMind CEO Demis Hassabis echoed concerns about misuse, citing the possibility of AI-enabled cyber attacks, bioweapons and election interference. One can only imagine the sense of alarm one might feel hearing this from the CEO of a leading AI research lab, which is rather like a pastry chef warning about the overwhelming power of cake.
Absent from the summit were representatives from many countries in the Global South, prompting some to raise the ever-cheerful topic of global inequality in technological governance. Because nothing says “international cooperation” quite like a panel of rich nations deciding on behalf of everyone else what the robot apocalypse should look like.
Still, organizers painted the summit as a success, largely on the basis that everyone showed up, behaved and signed something. In diplomatic terms, this is akin to a gold star on a behaviour chart.
So while it remains unclear whether the AI optimists are steering us toward utopia or simply letting the algorithms take the wheel and hope for the best, at least we can rest assured the world’s leaders have reminded us they know how to hold a nicely catered conference.
Because when it comes to saving humanity from superintelligent machines, nothing beats a well-worded memo.