The IMF report on AI states that it creates wealth. Il Manifesto carries out reverse fact-checking

There is a legitimate way to read a report: selecting the most important passages, highlighting the risks, and challenging its conclusions. And then there is another way: taking a document and presenting it almost in reverse. This is what happened on the front page of Il Manifesto, where a report by the International Monetary Fund on artificial intelligence is summarised as follows: “It does not create wealth; it helps to concentrate it in the hands of a few. An IMF report takes stock of AI: it increases inequality and devastates the planet. And data centres are also a threat of war.” It is a pity that the report states the opposite in several passages. Let us start from the beginning. AI, the Fund writes, “has the potential to boost productivity and growth”. Not exactly the definition of a technology that “does not create wealth”. For Europe, the cumulative gain in productivity could be around 1 per cent over five years; in a more favourable scenario, globally, GDP could be almost 4 per cent higher over a ten-year period. The report speaks of “productivity gains”, of benefits to be “maximised”, and of “stronger, more inclusive and sustainable” growth. One can debate every figure. But to claim that the Fund is saying AI does not create wealth is to turn a warning about distributional risks into a denial of the economic benefits set out in black and white.

The same pattern can be seen with regard to inequalities. The Fund does indeed warn that AI may exacerbate some of them, that part of income may shift from labour to capital, and that benefits may become concentrated. But it also adds that AI is “widening some gaps and narrowing others”, that the net effect on employment remains uncertain, and that around half of the jobs most at risk could become more productive thanks to their complementarity with AI. Even more curious is the claim that AI “devastates the planet”. There is nothing of the sort in the document. Instead, there is a substantive discussion on data centres: they currently consume around 3 per cent of Europe’s electricity; demand will grow; more networks, more interconnections and more low-emission energy will be needed. And then there is the gem about data centres being a ‘threat of war’. In the report, the geopolitical reasoning goes in the opposite direction. Europe is too dependent on the United States and China for chips, models and technological capabilities; to reduce this vulnerability, the Fund suggests strengthening Europe’s own capabilities in data centres, networks and semiconductor supply chains. In short, data centres are presented as a key component of Europe’s strategic resilience. Finally, there is a misunderstanding that needs clarifying.

The ‘international moratorium’ referred to on the front page is not a proposal by the International Monetary Fund: it is the position of the interviewee, David Krueger. The word ‘moratorium’ does not appear in the report. On the contrary, the Fund insists on regulation capable of reducing risks without stifling incentives to innovate and adopt advanced technologies. The controversy therefore does not lie in denying the risks of AI. These risks do exist; the Fund lists them. The controversy lies in the method: selecting all the risks, omitting almost all the benefits, and then presenting the result as ‘the IMF report’. It is a highly creative way of carrying out reverse fact-checking. The document states: AI can boost growth and productivity, but its costs and imbalances must be managed. *Il Manifesto* translates this as: it does not create wealth, it increases inequality, it devastates the planet. Rather than a summary, it seems like an unwitting refutation of the very source it purports to summarise.