As OpenAI Unveils Big Update, Protesters Call for Pause in Risky 'Frontier' Tech

About a dozen protesters rallied outside OpenAI’s San Francisco headquarters on Monday morning, calling for a measured pause in the development of the next generation of artificial intelligence ahead of a global policy meeting in Seoul.

The Microsoft-backed maker of ChatGPT is expected to take part in next week’s AI Seoul Summit, where industry leaders will discuss commitments made last year in a declaration on AI risks that promised technology transparency and human oversight to mitigate human rights impacts, privacy harms and unwanted bias in AI models.

The protesters outside OpenAI urged engineers to go farther in lessening the risks of their technology, carrying signs that read “Quit your job at OpenAI. Trust your conscience,” and “When in doubt, pause.”

At the same time, OpenAI announced in a YouTube livecast a more advanced version of its large language model and chatbot, which it touted as “a step towards much more natural human-computer interaction.” The updates boast faster response times and new audio and video capabilities — including the ability to read a person’s mood from their face — for the flagship product from OpenAI, which is valued at more than $80 billion, according to CNBC.

Representatives for the company did not return a request for comment.

Liron Shapira, a spokesperson for PauseAI, said the nonprofit group hopes regulators at the Seoul summit set a precedent that companies don’t have the right to unilaterally unleash increasingly advanced artificial intelligence.

“We think that we’re entering a regime of AI capabilities that nobody understands and nobody knows how to control, and there’s no turning back once we get there,” Shapira said. “And so we’re advocating for more research and more understanding before we plow ahead.”

PauseAI is advocating for a pause in the development of advanced AI systems that are more powerful than ChatGPT, called “frontier models,” which it argues could become a threat to humanity. The organization is not asking for a pause in more routine AI products.

“There’s only a few more steps of progress that these AI companies can take in making the AI smarter before it basically exceeds the intelligence level of humanity, which is something that nobody knows how to control,” Shapira said.

PauseAI is not alone in its concerns. This year, California lawmakers have advanced roughly 30 AI-related bills, some aimed at safeguarding the public, that may be seen as legal landmarks. And the declaration published at last November’s summit at Britain’s Bletchley Park, which commits world leaders to mitigating cybersecurity risks, was signed by 28 governments.

“There is potential for serious, even catastrophic, harm, either deliberate or unintentional, stemming from the most significant capabilities of these AI models,” the agreement notes. “Given the rapid and uncertain rate of change of AI, and in the context of the acceleration of investment in technology, we affirm that deepening our understanding of these potential risks and of actions to address them is especially urgent.”

Among those who signed the policy paper, which also noted AI systems have the potential to “transform and enhance human wellbeing, peace and prosperity,” were representatives of the United States, Japan, China and the Republic of Korea.

However, next week’s Seoul summit has far fewer attendees than the previous meetup in Bletchley Park, according to Reuters, with governments like Brazil and companies like Mozilla opting out, raising concerns from PauseAI that the gathering will fall flat.

“We need an international treaty,” PauseAI founder Joep Meindertsma said in a statement. “The 22 people inside that room in Seoul need to realize that they are the only ones who have the power to stop this race.”