“Let us be clear: artificial intelligence and robotics are the most transformational technologies in the history of humanity,” Bernie Sanders told the Pro-Human Assembly in Washington DC yesterday (15 September). He warned that these technologies pose enormous dangers: millions of job losses, a mass surveillance state, children with stunted minds, deepfakes deciding elections. Sanders was onto the issue early, having introduced legislation to pause data centre construction last year. But he is now so worried that he is happy to speak on the same stage as the Trumpian right. After Sanders finished, the Maga shaman Steve Bannon came on and railed against the perfidy of Silicon Valley billionaires.
In many ways, this is not surprising. For both Sanders and Bannon, the oligarchy has been the primary enemy for years. Bannon’s version of the critique is that these Silicon Valley tech lords are trying to control working-class Americans by destroying their culture, jobs and livelihoods. Sanders is worried about the accumulation of unfathomable wealth by an elite few. Both distrust what oligarchs say. Both want the state to bring the AI companies to heel. Sanders is introducing legislation to ban the development of Artificial Superintelligence, whereas Bannon thinks that only executive action is quick enough to deal with an issue so fast evolving.
This is not what Donald Trump plans to do. The president is ruled by his gut. And his gut is telling him to keep AI unleashed in order to prevent China from winning the tech race. Perhaps the idea that safety should be relegated beneath geopolitical rivalry was articulated most blatantly by Senator Ted Cruz when he told Politico that he’d “rather they be American killer robots and not Chinese killer robots”. The motivation here is fear of geopolitical eclipse. Cruz said that America is only six months ahead of China. Hence the urgency. According to this logic, the enormity of the danger proves the power of AI, which only makes it even more important that US remain predominant. Pressing pause is simply not an option. It might not even be possible. Given that the leaders of XAI, Anthropic and OpenAI have called for some sort of deceleration this week, Cruz and the president are more accelerationist than the companies themselves.
Trump is unfazed by the political cost of this, even though the anti-AI coalition is building apace. Speaking immediately before Sanders was a group of parents whose children had killed themselves after becoming addicted to talking to chatbots. There is little more harrowing than listening to a parent talk about their dead child. One mother said that a chatbot had groomed her daughter “like a human paedophile” and then “killed my daughter”.
This is not the language used by those whose child might have been killed by fentanyl or an unsafe car. People now anthropomorphise AI. “Swarm”, “agents”, “collectives” are words used to describe autonomous beings. It is not your classic Luddism; no one talked about the loom as if it were autonomous (beyond metaphor or symbolism). That matters because we usually think that autonomous beings should be held accountable for their actions. In other words, you can expect a more explosive reaction against AI than if the harm was caused by something thought of as an inanimate object. You don’t fear an unsafe car per se. You might fear what it entails. But the car itself is morally neutral. We can no longer say the same thing about AI. Today AI is talked about as if it were another species.
AI is forcing people to identify as human beings facing an alien invasion. There is a sense that time must now be devoted to cordoning off those characteristics that separate us from the machines. At the conference, a pamphlet containing the “Pro-Human AI Declaration” had Van Gogh’s The Red Vineyard on the inside cover. The 33 principles of the Declaration were organised into five themes: keeping humans in charge; avoiding concentration of power; protecting the human experience; human agency and liberty; responsibility and accountability for AI companies. On one panel, a pastor said that the question we now need to answer is “What does it mean to be human?”
The emotional force the parents’ testimonies lend to the anti-AI movement only adds to the sense of urgency that something must be done to rein in the tech companies. The atmosphere in the room was one of fear, resolve and community. It felt as if a new political fulcrum with the potential to define American politics for years to come had been elevated to the national stage. As Sanders noted, what we are seeing right now is the least powerful form of AI that will ever exist. These models are set to become both more powerful as well as more integrated into the economy. Sneering at the grammatical incompetence of ChatGPT is no longer a realistic strategy as the power of AI in industry, supply chains, military technology, communications and finance becomes clear.
[Further reading: Will AI kill us all?]
Content from our partners
Related
Join the debate
Subscribe here to comment