Jack Clark, co-founder and head of policy at AI firm Anthropic, has predicted that artificial intelligence could begin contributing to scientific discoveries worthy of major prizes within the coming months, potentially leading to a Nobel Prize nomination within the next year. Speaking at Oxford University, Clark described the current pace of AI development as creating a “vertiginous sense of progress,” with breakthroughs arriving far faster than many expected.
Rapid Advancements and Bold Predictions
According to Clark, an AI system working alongside humans could help make a Nobel Prize-winning discovery within 12 months. He also anticipates that bipedal robots will begin assisting tradespeople within the next two years, while businesses run exclusively by AI could be earning millions of dollars in profit within 18 months. Furthermore, Clark predicts that by the end of 2028, AI systems will be designing their own successors.
Clark’s statements echo bold predictions from leaders in the artificial intelligence industry. Anthropic, founded by former OpenAI scientists concerned about AI safety, has become a prominent player in the field. The company focuses on developing “reliable, interpretable, and steerable” AI systems, and its chatbot Claude competes with OpenAI’s ChatGPT and Google’s Gemini.
Warnings and Comparisons
Despite his optimism, Clark warned that many people underestimate how rapidly AI systems are advancing. He compared society’s underpreparedness for AI threats to the lack of preparedness for the COVID-19 pandemic, stressing that governments risk being reactive rather than proactive if they fail to foresee the implications of AI technologies.
Clark also highlighted potential dangers, stating there is a “non-zero chance” that AI could one day turn against humanity. He suggested slowing development to better understand its impact, but acknowledged this is unlikely due to intense competition among nations and businesses.
Broader Implications and Concerns
Clark’s remarks come amid increasing international debate over AI regulation and safety. Anthropic has positioned itself as a conservative player in the field, with some former OpenAI staff advocating for stricter controls on frontier AI systems to prevent abuse, misinformation, and autonomous actions.
During the lecture, Clark mentioned a recently developed system named Mythos, which shows promise in identifying cybersecurity flaws. He also predicted that AI could eventually lead to scientific progress with minimal human involvement and create forms of scientific equipment not yet imagined.
Concerns Over Human Dependence
Beyond existential threats, researchers are discussing the societal and psychological effects of extensive reliance on AI. Edward Harcourt, director of the Institute for Ethics in AI at Oxford University, warned during the same event that humanity could suffer from “cognitive atrophy” if AI takes over too much human cognition. He called for “Socratic” AI systems that foster reasoning rather than replacing humans entirely.
Clark conceded that some of his forecasts might be exaggerated, but insisted that an unprecedented revolution is inevitable. As governments, researchers, and tech firms race to shape AI’s future, his remarks underscore the sharp divide between excitement over rapid innovation and fear of losing control over increasingly powerful systems.
