Risk of extinction in exchange for living much better: the provocative calculation of the new economist of an AI giant

The phrase seems taken from a science fiction novel. But it was written by Chad Jonesthe economist who has just joined Anthropicone of the companies artificial intelligence most important in the world.

Jones is a professor of economics at Stanford University and specializes in studying How technology influences economic growth. In recent years, he has focused much of his work on artificial intelligence and the transformations it could trigger.

As highlighted by Financial Times (FT)the academic maintains a vision markedly optimistic about the potential of AI, something that made him an influential voice within the circles that analyze the development of this technology.

The controversial phrase appears in an academic work published in 2023 by the United States National Bureau of Economic Research (NBER), titled “The AI ​​dilemma: growth versus existential risk”.

As a theoretical exercise, Jones proposed a scenario in which artificial intelligence generates a risk of extinction of humanity of the 1% annually over a period of 40 years.

Under that assumption, the probability of civilization surviving At that stage it would be approximately two thirds.

From these numbers, the economist concludes that it would be “optimum” accept a one in three chance of the disappearance of the human species in exchange for a two-thirds chance multiply living standards by 55.

The statement does not constitute a prediction about the future or a warning that this scenario will occur. It is a economic modeling exercise that tries to put into numbers the dilemma between the potential benefits of artificial intelligence and its most extreme risks.

The incorporation of Jones also draws attention for the place he reaches. Anthropic presents itself as one of the most risk-conscious AI companies associated with the development of this technology.

Anthropic’s main product is Claudea family of generative artificial intelligence models capable of maintaining conversations, writing texts, analyzing documents and assisting in programming and professional work tasks. The company competes directly with OpenAIcreator of ChatGPT, in the race to develop increasingly advanced AI assistants.

Since its launch, Anthropic tried to differentiate itself by putting a strong emphasis on safety and in the responsible development of artificial intelligence. That approach made Claude one of the most used platforms by companies and developers looking to incorporate generative AI tools into their work processes.

The co-founder and CEO of Anthropic, Dario Amodeihas long argued that both the opportunities and dangers of AI are being underrated.

In an essay published in 2024, he stated that “most people are underestimating the extent to which the benefits of artificial intelligence could be revolutionary, just as they are underestimating how serious your risks could be“.

The arrival of jones seems to align with that vision. The economist is among those who believe that artificial intelligence could produce economic transformations of unprecedented magnitude, although his own work shows that these scenarios also raise questions that are difficult to answer.

In a recent interview with the Financial Times, Amodei summed up his view with a phrase as optimistic as it was enigmatic: “There is no end on the other side of the rainbow. There is only the rainbow.”