Crashing waves vs. rising tides: Overturning prior views about how AI could overtake human workers

In recent years, the rapid advancements of artificial intelligence (AI) have sparked debates about its potential to overtake human workers in various fields. Anthropic CEO Dario Amodei has predicted that AI could surpass "almost all humans at almost everything" shortly after 2027. While AI's capabilities are undeniably improving, such rapid progress might seem at odds with existing evidence. Studies have shown that AI is still failing at 95%+ of remote freelance projects and continues to struggle with hallucination, long-term planning, and forms of abstract reasoning that humans find easy. However, recent research from the Machine Intelligence Research (MIRI) organization, now known as METR, has found evidence that large language models (LLMs) can gain capabilities in rapid surges—jumping from succeeding almost never to almost always in just a few years. If this trend holds true across the economy, it could mean that workers could be blindsided by AI advances.

The initial skepticism about AI's potential to outperform humans stems from its current limitations. AI systems often struggle with tasks that require common sense, creativity, and the ability to think abstractly. For instance, when given a task like writing a compelling blog post or solving a complex math problem, AI models frequently produce inaccurate or nonsensical responses. These shortcomings have led many experts to believe that AI would not be able to replace humans in the near future.

However, the METR study challenges this view by highlighting the potential for sudden, significant improvements in AI capabilities. The researchers observed that LLMs can experience rapid surges in performance, often within a few years. This phenomenon could mean that AI systems might leapfrog over human abilities in certain domains more quickly than previously anticipated.

If this trend continues, the implications for the workforce are profound. Workers in industries reliant on tasks that can be automated might find themselves displaced by AI systems that can perform the same tasks more efficiently and at a lower cost. This could lead to widespread unemployment and economic disruption.

But the potential for AI to overtake humans is not limited to low-skilled jobs. The METR study suggests that even complex, high-level tasks could see rapid advancements. For example, AI might improve its ability to diagnose medical conditions, develop new drugs, or even compose original music. In these cases, AI could potentially replace human experts, leading to significant changes in the job market and the nature of work.

The challenge for policymakers and businesses lies in preparing for this potential shift. Governments will need to develop strategies to mitigate the negative impacts of AI-driven job displacement, such as reskilling programs and safety nets for displaced workers. Businesses, in turn, must invest in AI research and development to stay competitive, while also ensuring that their operations remain adaptable to changing technological landscapes.

In conclusion, while the idea of AI surpassing humans in nearly every domain might seem far-fetched, recent evidence from METR suggests that rapid surges in AI capabilities are possible. If this trend continues, the workforce and economy could face unprecedented challenges. It is crucial for stakeholders to remain vigilant and proactive in preparing for a future where AI plays a more significant role in shaping the labor market.