Artificial intelligence is not just another tool—it’s a game-changer that is already reshaping the economy, politics, and our behavior. The exponential progress of recent years has meant that we’ve been living with it for 16 years without fully understanding it. If we fail to manage it, we will pay with a loss of attention, relationships, and our ability to make autonomous decisions.
Exponential change, not just another tool
A car cannot make a better car and a reactor will not build a new reactor, but artificial intelligence is already developing further artificial intelligence. Its capacity has increased from the first models to today’s systems by roughly ten-thousandfold—a curve we find hard to perceive because it is exponential and measured on a logarithmic axis. The race for general artificial intelligence now drives global politics, especially between the USA and China. What matters, however, is that the consequences arrive sooner than we can process them.
Power is also shifting: while a decade ago data were largely in the public sector, today they are overwhelmingly in the hands of private companies. In the economy we see a leap in productivity, but also deep restructuring and a surplus of “cognitive capacity” that is no longer the exclusive domain of humans. Work will change significantly, though this need not mean a mass loss of jobs. Geopolitics is again shifting toward a bipolar world, with Europe still searching for a clear trajectory.
The human cost: skill atrophy and fragile relationships
Our “human agency” is weakening: we live in information bubbles, and we succumb more quickly to manipulation and polarization. Among young people, mental health disorders and loneliness are surging—loneliness that the U.S. Surgeon General compares to smoking 15 cigarettes a day. Older, “analog” generations are more resilient to these influences because they did not start out online. This shift also has physical consequences, because mental stress and isolation lay the groundwork for disease.
There is a risk of atrophy in key skills—empathy, compassion, critical thinking, and attention. Early experiments (even though they have not yet undergone full peer review) suggest a risk of long-term cognitive atrophy when relying on generative tools. The impact is cumulative: an hour means nothing, but years spent across various applications change habits as well as identity. There is also a rise in “greenhouse” relationships with digital partners—enticing, always pleasant, yet they make us fragile in the real world; after all, a portion of young men in the USA do not enter real relationships at all.
How to prepare: policy, workplaces, resilience
We need to triple investment in AI safety, create a European “CERN” for artificial intelligence, retain talent, and build an alliance of democratic countries. Europe is not lagging because of regulation, but because of comparatively low investment; without resources we will not become pioneers. Equally important are communities—family, civic, workplace, and educational—that cushion loneliness and strengthen the sense of belonging. Policies must reflect the exponential dynamics of technologies, otherwise they will respond too late.
In companies, two-thirds of success comes from people and only one-third from tools, even though most organizations are already experimenting with AI and truly successful ones are still few. Individuals and teams should deliberately train: emotional control, flexible thinking, a strong “self,” confidence in action, and a sense of connection with others. Schools and workplaces should include technical self-confidence, attention control, emotional autonomy, the cultivation of deep relationships, and compassion. The skills of the AI era—critical and ethical thinking, emotional intelligence, intercultural communication, creativity, and above all flexibility—are our best insurance for preserving both our humanity and our autonomy.
