Dn the economic debate, there are those who assure that any disruptive innovation is good for employment, and history has so far proven them right: steam, electricity and computing have not created of cohorts of unemployed over the last two centuries, quite the contrary. And then there are those who assure that generative artificial intelligence (AI), which is rapidly spreading from industry to services via agriculture, is on the contrary a mortal threat for many professions.
Without ignoring the beneficial contributions of AI – on productivity in particular – researchers from the International Monetary Fund (IMF) warn against the side effects of the innovation popularized by ChatGPT (OpenAI-Microsoft) or Gemini (Google) .
In January, the general director of the institution, Kristalina Georgieva, already warned that “nearly 40% of jobs are exposed to AI” and even 60% in advanced countries. “This revolution that could raise incomes around the world also risks replacing jobs and widening inequalities”as planned “most scenarios”worried Kristalina Georgieva.
“Readiness Indicator”
The IMF therefore launched, in 2023, an “AI preparedness indicator” to measure in particular the level of social protection in the face of such risks. Its researchers emphasize that an active tax policy in this sense has “a major role to play” – including in favor of graduates – in order to “support a more equitable distribution of gains and opportunities linked to generative AI”.
Governments must therefore find new resources, without taxing AI itself so as not to slow down its benefits. The authors of the note invite them, on the other hand, to increase the taxation of capital to reduce inequalities, to increase unemployment benefits and to invest in training to prepare “to jobs in the AI era”.
The IMF’s point is of course general: the welfare state is non-existent in many countries, often in crisis in others. In France, more than the concern for protection against the risks linked to AI, it is the search for full employment and the major injunction from Emmanuel Macron to ” work more ” which dominated political choices and justified the unemployment insurance reform.
So far, ChatGPT has not led to the end of work, and nine out of ten current jobs already existed a century ago. But the question remains: is the “technological unemployment” described by the economist John Maynard Keynes a necessary transition between two upheavals in the productive system or a reality destined to last?