Jack Clark makes prediction in Oxford address
Pro
Jack Clark, Anthropic
Jack Clark, co-founder of Anthropic, believes that the evolution of AI is progressing at a dizzying pace and predicts that within a year a Nobel Prize-level breakthrough will be achieved through collaboration between humans and AI. During a presentation at the University of Oxford reported on by The Guardian, Clark sketched a future in which bipedal robots will be assisting workers within two years and AI-led enterprises will be generating millions in revenue within 18 months. He went on to suggest that by the end of 2028 AI could have the ability to develop subsequent versions of itself.
Despite this progress, Clark stressed that the existential threats posed by AI remain significant, pointing to the real possibility that the technology could lead to the extinction of humanity.
He underlined the dangers of current capabilities, citing Anthropic’s Mythos model, which demonstrated a disturbing ability to circumvent cyber security measures. Although he conceded that slowing development would give humanity more time to adapt, he argued that geopolitical and commercial competition between countries and companies makes such a pause unlikely, as the race for dominance outweighs global safety concerns.
Anthropic was originally founded by former OpenAI staff who prioritised safety protocols. While some critics and politicians have accused the company of using safety warnings to manipulate regulation and gain a market advantage, Clark maintains that society is largely in denial about the scale of AI’s growth.
He warned that failing to prepare proactively for a superintelligence – which may soon surpass humanity’s collective capabilities – would mirror the catastrophic lack of preparedness seen during the Covid-19 pandemic, leaving humanity in a permanent state of reaction.
Systemic vulnerabilities
The concentration of power in a few profit-driven entities such as Google, OpenAI and Anthropic is fuelling concerns about systemic vulnerabilities. In addition, Professor Edward Harcourt, head of the Institute for Ethics in AI, warned that relying on AI to carry out complex tasks could lead to “cognitive atrophy”, eroding human judgement. He proposed the introduction of ‘Socratic AI’, designed to challenge and stimulate human intellect rather than replace it.
Concluding his speech, Clark noted that even his most cautious predictions imply a fundamental transformation of the world economy and social structures. He envisaged a future in which scientific discoveries take place independently of humans and a “machine economy” emerges, detached from human financial systems – admitting that these possibilities may sound far-fetched, but that they reflect the profound shifts currently under way.
Emerce