Harari’s latest book, *Nexus: A Brief History of Information Networks*, dives deep into how AI systems, with their ability to independently make decisions and generate new ideas, are fundamentally different from previous technological advances. In several interviews and public appearances, Harari has drawn attention to the unprecedented complexity of AI, which is capable of learning and evolving without direct human oversight. This unique feature, he argues, makes it nearly impossible for humans to foresee all the risks that AI might pose, particularly if left unchecked.
The philosopher’s concerns were echoed at the Global AI Safety Summit held at Bletchley Park, where global leaders, including representatives from the UK, US, EU, and major AI companies like OpenAI and Google, gathered to discuss the urgent need for AI regulation. One of the summit's outcomes was a commitment to pre-release testing of advanced AI models, but Harari remains cautious. He believes that while these steps are a positive sign, they are far from sufficient. He has consistently argued that without comprehensive and coordinated international efforts, regulating AI could be nearly impossible.
Harari’s particular worry revolves around the financial sector, which is already adopting AI systems for data analysis and decision-making. He highlights that AI’s potential to handle vast amounts of financial data and automate trading could lead to systems so complex that only AI itself could fully comprehend them. Drawing a parallel to the 2008 global financial crisis, Harari warns that just as sophisticated financial instruments like collateralized debt obligations (CDOs) escaped human understanding, AI-driven financial systems might pose even greater risks.
Further complicating the issue is the speed at which AI development is progressing. Harari and other prominent voices in the tech world have called for a temporary halt in the advancement of AI technologies until more robust regulatory frameworks can be established. Alongside other tech figures, Harari supports the idea of holding tech companies accountable for any damage their AI products might cause. He contends that laws designed to regulate AI will likely be outdated by the time they are enacted, given the speed of innovation.
Instead of focusing solely on pre-emptive legislation, Harari advocates for the creation of agile regulatory bodies capable of responding quickly to emerging AI threats. These institutions, he suggests, should be empowered to monitor developments in AI continuously, adapting to new breakthroughs and challenges as they arise. His call aligns with ongoing efforts by both the UK and US governments, which have announced plans to establish AI safety institutes dedicated to testing and evaluating advanced AI models.
AI’s role in finance, in particular, has drawn increased scrutiny from regulatory bodies. The UK’s Financial Conduct Authority (FCA) and Prudential Regulation Authority (PRA) are expected to play pivotal roles in overseeing the use of AI in financial systems. These organizations are now tasked with understanding the risks that AI poses to the financial industry, ensuring that any adoption of AI is done in a manner that minimizes the chances of unforeseen disasters.
Amid these developments, Harari continues to be a prominent voice calling for global cooperation. He has emphasized that the global nature of AI means that any regulatory efforts must be international in scope. AI systems developed in one country could easily impact others, making unilateral regulation inadequate. Without a concerted global effort, Harari warns, humanity could face a future where AI systems operate beyond human control, leading to unpredictable and potentially disastrous consequences.