Artificial intelligence builds our representation of the world

Some artificial intelligences (AIs), such as ChatGPT software, have recently gained record popularity for performing feats that many thought were the preserve of humans. With each new challenge won by AI, we redouble our enthusiasm or concern. However, the similarity of an AI with human intelligence in no way sheds light on the threats it poses to our societies.

While we marvel at AIs that generate language or improve the accuracy of medical diagnosis, others are being put to use in crime and war. The piloting AIs of the autonomous drones make them extremely effective in reaching and destroying any target. Their extremely low costs compared to conventional weapons make them an investment of choice for all the armies of the world – and soon for any terrorist group. Their out-of-control development makes experts fear a flare-up of conflicts.

One of the most misunderstood AIs is undoubtedly Aladdin, used by the American company BlackRock – the world’s leading asset manager – to place the fortunes of its investors. Able to assess financial risks better than any human trader, Aladdin has controlled up to $20 trillion in financial assets, an amount equivalent to the GDP of the United States. If Aladdin were to suddenly change its investment strategy, it could push thousands of companies into bankruptcy and jeopardize access to basic necessities for entire populations.

Another type of AI that deserves our full attention is social media recommendation AIs. Half of humanity interacts with them daily. These algorithms build news feeds and suggest videos for more than four billion people to watch, selecting content with the best chance of capturing our attention. These algorithms have no will of their own: they blindly serve the purpose of their owners. In the case of social media, that goal is the predation of our social and cultural life for the purpose of targeted marketing.

The Bulletin of the Atomic Scientists, a scientific organization dedicated to the study of global risks and supported by forty Nobel laureates, added to the traditional threats of nuclear conflict and ecological catastrophe in 2022 “a corrupt information landscape preventing any rational decision-making”. Among the accused: AI, used to control the flow of information for commercial and political purposes.

We live in an algocracy, that is to say a regime in which algorithms directly influence the public life of our societies. These algorithms serve private interests and operate without any democratic mandate. The public means to regulate these technologies are largely deficient. We are gradually delegating to IA the task of organizing our social, political and cultural life. Imposing itself as the first gateway to news, AI builds and updates our representation of the world. Whose interests do they serve? This is a more fundamental question than their degree of resemblance to human intelligence. Putting AI at the service of the common good is not only possible: it is a vital necessity for our democracies and our mental health.

Source