The excitement around the London arrival of OpenAI CEO Sam Altman was palpable from the queue that snaked its way around the University College London building ahead of his speech on Wednesday afternoon. Hundreds of eager-faced students and admirers of OpenAI’s chatbot ChatGPT had come here to watch the UK leg of Altman’s world tour, where he expects to travel to around 17 cities. This week, he has already visited Paris and Warsaw. Last week he was in Lagos. Next, he’s on to Munich.
But the queue was soundtracked by a small group of people who had traveled to loudly express their anxiety that AI is advancing too fast. “Sam Altman is willing to bet humanity on the hope of some sort of transhumanist utopia,” one protester shouted into a megaphone. Ben, another protester, who declined to share his surname in case it affects his job prospects, was also worried. “We’re particularly concerned about the development of future AI models which might be existentially dangerous for the human race.”
Speaking to a packed auditorium of close to 1,000 people, Altman seemed unphased. Wearing a sharp blue suit with green patterned socks, he talked in clipped answers, always to the point. And his tone was optimistic, as he explained how he thinks AI could reinvigorate the economy. “I’m excited that this technology can bring the missing productivity gains of the last few decades back,” he said. But, while he didn’t mention the protests outside, he did admit to concerns over how generative AI could be used to spread disinformation.
“Humans are already good at making disinformation, and maybe the GPT models make it easier. But that’s not the thing I’m afraid of,” he said. “I think one thing that will be different [with AI] is the interactive, personalized, persuasive ability of these systems.”
Although OpenAI plans to build in ways to make ChatGPT refuses to spread disinformation, and plans to create monitoring systems, he said, it will be difficult to mitigate these impacts when the company releases open-source models to the public—as it announced several weeks ago that it intends to do. “The OpenAI techniques of what we can do on our own systems won’t work the same.”
Despite that warning, Altman said it’s important that artificial intelligence not be overregulated while the technology is still emerging. The European Parliament is currently debating legislation called the AI Act, new rules that would shape the way companies can develop such models and might create an AI office to oversee compliance. The UK, however, has decided to spread responsibility for AI between different regulators, including those covering human rights, health and safety, and competition, instead of creating a dedicated oversight body.
“I think it’s important to get the balance right here,” Altman said, alluding to debates now taking place among policymakers around the world about how to build rules for AI that protect societies from potential harm without curbing innovation. “The right answer is probably something between the traditional European-UK approach and the traditional US approach,” Altman said. “I hope we can all get it right together this time.”