WritingResources
Back to all posts

Will the AI Elite let the public speak?

Ludivine Siau

In short

The gap is widening between the public’s concerns about AI and the industry’s push for rapid and unregulated development. Recent reports are showing the public's uneasiness with the current pace of AI development and their lack of support for the quest for superintelligence until systems are proven safe. This gap was recently highlighted when a spat with Anthropic exposed how the US government and its allies view safety measures and regulation as a threat to their acceleration agenda.

Quick breakdown

The public’s demand for caution is finally emerging, with concerns about the “move fast and break things” approach of the AI industry. A report from the Pew Research Center earlier this year indicated that while the public believes in the potential benefits of AI, they are worried that government regulation will be too lax. This seems to align with what the Future of Life Institute reports from their survey of 2,000 US adults:

  • Only 5% of respondents support the status quo of fast, unregulated development
  • 73% support robust regulation
  • 64% feel that superhuman AI should not be developed until proven safe and controllable, or should never be developed

The organisation also started a petition, “Statement on Superintelligence”, which calls for a ban on developing superintelligence until proven safe and accepted by the public. It has already gathered over 41,600, including those of influential researchers, business figures and policymakers.

Meanwhile, there is a tense dynamic at play between the AI industry and authorities. The largest AI players–the “hyperscalers”–are doing everything they can to play nice with the government, Congress, and states to keep the road unblocked for their growth. This strategy is highly effective, but fragile, as shown when Anthropic co-founder Jack Clark argued for appropriate caution, transparency and public consultation in a recent blog post. Clark very reasonably stressed that AI’s complexity means it cannot yet be fully trusted without oversight. This pragmatic position was quickly attacked by David Sacks, a Trump administration AI advisor and prominent Silicon Valley VC, who accused the AI lab of promoting regulation that would work in their benefit and harm AI startups. As a response, Anthropic CEO Dario Amodei released a statement reaffirming the company’s alignment with the US government’s AI leadership. Crisis aborted and lesson learned: you can’t flatter the King enough, these days.

While deregulation is the rule of the game at federal level, there is hope for more sensible approaches at state level:

  • California enacted new legislation for transparency in frontier AI and for the safety of chatbots.
  • New York became the first state to ban AI-enabled rent price fixing.

Why you should care

The public is moving ahead of authorities in grasping the stakes of unchecked AI development. Despite the technology vastly escaping popular understanding, citizens are intuitively demanding transparency and guardrails. They are not rejecting AI outright. Rather, they are acting as sensible consumers who believe in the technology's potential and are aware that the future is also fraught with risks.

This US-centric policy discussion holds great importance across borders because the regulation where the technology is developed (or, in this case, the lack thereof) often sets the global standard. Sadly, as demonstrated by the backlash against Anthropic's cautious tone, the top of the pyramid views even the most basic safety measures as acts of sabotage or political betrayal.

We can only hope that a change in direction will come from the ground and local efforts, which can then ripple outward. For instance, the sensible efforts made by California to regulate transparency and chatbot safety might have an outsized impact, given that it is at the heart of the AI ecosystem.

Sources