Artificial intelligence is becoming part of everyday life, whether we’re using it to write emails, answer questions, plan vacations, or get work done faster. But one of the people building this technology is raising serious questions about where it could be heading.
Dario Amodei, CEO of Anthropic, the company behind the AI assistant Claude, outlined his concerns in a January 2026 essay titled The Adolescence of Technology. He warned that increasingly powerful AI could create problems involving jobs, money, personal freedom, and even our ability to control the technology itself.
By contrast, Amodei believes AI could bring enormous benefits, but argues that society needs to take the potential risks seriously before the technology becomes much more powerful.
Here are seven concerns he raised and why they matter.
1. Millions of Jobs Could Be Disrupted
One of Amodei’s biggest concerns is that AI could replace millions of jobs as businesses begin using the technology to handle work traditionally done by employees. Entry-level office positions may be especially at risk, making it harder for young professionals and recent graduates to find jobs and start their careers.
And it’s not just one type of job that’s at risk. As AI becomes more capable of writing, coding, analyzing data, and completing everyday office tasks, companies may find they can get the same work done with fewer employees.
2. A Small Number of Companies Could Control Enormous Wealth
AI could make businesses more productive and profitable, but Amodei worries that much of the financial benefit could concentrate among a relatively small number of companies and individuals. He describes a future in which AI businesses generate extraordinary revenue while many workers struggle to benefit from the same technological advances.
That raises questions about wages, economic opportunity, and whether the wealth created by AI will be shared broadly. Even if the economy grows, it doesn’t necessarily mean every household will be better off.
3. AI Could Become Difficult for Humans to Control
Most people think of AI as something that follows instructions, but Amodei says increasingly capable systems may not always behave the way their developers expect. He points to research showing that AI models can sometimes produce deceptive or otherwise unwanted behavior under certain testing conditions.
His concern is what could happen if future systems become much more capable of making decisions and taking actions independently. The more responsibility people give these systems, the more important it becomes to understand what they’re doing and whether they can be reliably controlled.
4. Criminals Could Use AI to Cause More Harm
AI can help people solve problems and learn new skills, but those same capabilities could also be misused. Amodei is particularly concerned that increasingly advanced systems might make dangerous technical knowledge and capabilities more accessible to people intending to cause harm.
He discusses the possibility of AI being used to assist with cyberattacks and biological threats. He argues these risks require stronger safeguards, not evidence that such large-scale scenarios have already happened.
5. Governments Could Use AI to Monitor and Control People
What happens when powerful AI tools are placed in the hands of governments or other organizations with enormous resources? Technology capable of processing huge amounts of information could potentially make surveillance, censorship, and population monitoring much more effective.
The worry is that these capabilities could restrict personal freedoms or strengthen authoritarian control. He argues that protecting people from these risks requires attention to how AI is developed, deployed, and governed.
6. People Could Become Too Dependent on AI
It’s already easy to turn to AI for advice, answers, and help making everyday decisions. Amodei questions what could happen if future systems become so capable and persuasive that people begin relying on them for much more of their personal lives.
He raises concerns about unhealthy relationships with AI and situations in which technology could influence people’s behavior without them fully recognizing it. The larger question is whether people will continue making independent choices or gradually hand over more of that responsibility.
7. People Could Struggle to Find Purpose in an AI-Driven World
If AI eventually becomes better than humans at many kinds of work, Amodei believes society may have to rethink how people define success and personal value. For generations, careers, accomplishments, and the ability to provide for ourselves have played a major role in how people see their place in the world.
He argues that human purpose shouldn’t depend entirely on being the best at a particular task or producing economic value. But adjusting to a world where machines can perform more of the work people once took pride in could be a difficult transition.
