Anthropic: Does AI Really Threaten Human Jobs?

A Threat Perceived as Inevitable

In Silicon Valley, the idea that artificial intelligence could trigger an employment apocalypse has become a widely shared belief. Discussions on this topic are so pessimistic that a researcher specializing in societal impacts at Anthropic recently predicted an imminent recession and a collapse of careers before the benefits of AI manifest. Dario Amodei, CEO of Anthropic, even claimed that AI could become a general substitute for human labor, capable of performing all professional tasks in less than five years. These alarming perspectives are not isolated to Anthropic but resonate throughout the industry.

These dire predictions have plunged many workers into anxiety, fueling movements aimed at halting the construction of new data centers. This panic is exacerbated by the lack of a clear response from lawmakers, who have yet to formulate a coherent strategy to address these changes. Even economists, who had previously downplayed the immediate impact of AI on employment, are beginning to consider that this technology could transform our way of working in unprecedented ways.

The Call to Action from Alex Imas

Alex Imas, an economist at the University of Chicago, expressed his concerns during a recent conversation. He emphasized that our current tools for predicting the impact of AI on employment are largely inadequate. He calls for a mobilization of economists to collect essential data that could facilitate the planning of AI integration into the workforce.

Imas criticizes the current method of assessing a job's exposure to AI based on individual tasks. For example, the work of a real estate agent includes tasks like interviewing clients about their property preferences. Since 1998, the U.S. government has cataloged thousands of these tasks, and this data has been used by OpenAI to evaluate job exposure to AI. However, Imas argues that this approach is misleading and does not effectively predict job displacement.

Exposure to AI alone is not enough to determine whether a job is at risk. Imas explains that even if a job is fully automatable by AI, that does not necessarily mean it will be replaced, as implementation costs can be prohibitive. Furthermore, for the majority of jobs, the situation is more complex, making it difficult to predict when and how AI will actually impact these positions.

The Limits of AI Exposure

Consider the case of a dating app developer. With AI, they can accomplish in one day what previously took three days, thereby increasing their productivity. This increased efficiency could prompt the employer to hire more to meet rising demand, or conversely, to reduce staff if demand does not increase proportionally. This uncertainty underscores the importance of understanding price elasticity and market reactions to cost changes.

Imas points out that we have detailed data for everyday consumer products like cereals, thanks to collaborations with supermarkets. However, there is no similar data for professions like web developers or dietitians, which are also exposed to AI. He calls for a massive data collection effort, comparable to the Manhattan Project, to track these statistics on an economy-wide scale.

Although this data collection is costly and time-consuming, Imas is convinced it is essential. It would provide economists with a realistic view of the future of work in an AI-dominated world and enable policymakers to prepare appropriate strategies to manage this inevitable transition.

Towards Better Anticipation of the Future

The impact of AI on the labor market is a complex and multifaceted issue. The ability of AI to replace certain human tasks raises questions about the future of jobs and how businesses will need to adapt. The efficiency gains achieved through AI could transform entire sectors, but without a clear understanding of the underlying economic dynamics, decision-makers risk making ill-informed choices.

Collecting detailed economic data could provide valuable insights into how AI influences various sectors and how businesses can adapt to these changes. Ultimately, it is about preparing the labor market for a future where AI will play an increasingly important role, while ensuring that workers are not left behind in this transition.