Mo Gawdat of Google X: AI's Biggest Threat Lies in Ethics, Not Code

Mo Gawdat, former chief business officer at Google X, suggests viewing AI not as a separate threat but as an amplifier of problems society already has. In his account, algorithms have long interfered in people's everyday choices, and the next stage is a blow to employment, the economy and the rules for distributing power.

Algorithms and attention

Gawdat starts with a simple point: artificial intelligence is no longer somewhere in a lab. It runs in feeds, recommendations, ad systems and search, which is to say wherever it is decided what a person will see, read and buy. Even the choice of what to read often happens unconsciously: a model optimized for engagement makes it on the user's behalf. So the conversation about the future of AI, in his view, is late: the influence is already built into the everyday digital environment.

This leads to the main shift. The question is no longer whether AI can do useful things, but who sets its goals and what incentives lie behind its actions. If a system is trained to hold attention at any cost, it inevitably pushes toward polarization, dependence on content and emotional reactions. Here the technology is not an independent villain but a precise instrument of human priorities.

Work and the system

The second line of his argument is economics. Gawdat believes unemployment in some sectors could reach 50%, because AI automates not only routine work but a significant share of cognitive work. Functions that were considered safe not long ago are under threat: analytics, support, marketing, initial legal work, and some engineering and creative tasks. This does not necessarily mean an instant collapse, but it does mean a very harsh transition for millions of professionals.

• White-collar automation
• Pressure on wages in service roles
• Growing profits for infrastructure owners
• Weak readiness of states for a rapid transition

Hence the more radical thesis of a crisis of capitalism in its current form. If productivity grows while income concentrates in a narrow group of companies and investors, the system begins to break down politically and socially. The question is no longer whether AI will become more efficient than a human at a given task, but who will capture the benefit of that efficiency and how economic power will be redistributed.

Ethics as the bottleneck

Gawdat's harshest thesis is that the problem is not superintelligence as such, but the state of people themselves at the moment it appears. He describes it this way: humanity is bringing its own old defects to a historically powerful technology: greed, manipulation, a cult of growth and weak moral constraints. In this setup, AI accelerates not the best qualities of civilization but its distortions. The stronger the tool, the more dangerous the bad motives of its owner.

"The main problem is not artificial intelligence, but human ethics."

This changes the frame of the whole debate. Instead of the familiar argument over whether machines will "rebel," Gawdat proposes discussing more down-to-earth matters: development rules, corporate accountability, algorithmic transparency and new social contracts. Otherwise society will get not an autonomous evil from science fiction but a thoroughly human system that scales inequality, pressure on workers and control over attention faster than institutions can react.

What this means

Such statements do not give a ready forecast with dates, but they do capture a shift: the conversation about AI is moving away from excitement over new models toward questions of power, the distribution of money and the quality of human decisions. If ethics and regulation do not catch up with the technology, the main risks will come not from the machine itself but from the people who already control its goals.