Few Tech Giants Control AI — Study Warns Your Rights Are at Risk
A handful of companies now decide which AI you get, and on what terms.
A researcher in Poland, Marcin Marciniak, has published a 17-page paper arguing that advanced artificial intelligence is piling up in the hands of a small club of technology companies. These firms don't just make AI models. They also own the computer chips, the cloud servers, the data, the apps you use, and the app stores you download from. That combination, the paper says, quietly turns ordinary market power into something closer to private government — these companies end up setting the rules for how your basic rights actually work in daily life.
The evidence comes from real places. The paper looks at the spread of giant data centres in Ireland, the United Kingdom, the United States and Chile, where the buildings that power AI consume huge amounts of electricity and water while creating relatively few local jobs. It also examines studies of workplace automation and union bargaining, finding a pattern: the benefits of AI spread widely and cheaply, but the money and the decision-making stay concentrated at the top. Locals get the noise and the power bills; the profits go elsewhere.
Crucially, the paper is careful to say that bigness alone is not a rights violation. But it does create four practical problems: you become dependent on a service you can't easily leave, you can't see how decisions about you are made, switching to a rival is expensive or technically painful, and the social and environmental costs get pushed onto the public rather than the company. Your photos, your work habits, your medical data — all may sit inside systems you have no way to inspect.
The author's recommendations are concrete. Competition law alone isn't enough, he argues. Governments should also run rights-based impact assessments before letting AI systems spread, force companies to make their services work with rivals so you can actually switch, fund public research so there's an alternative to private AI, give workers a real voice in how automation is deployed, and guarantee people a way to challenge decisions and get a remedy. In short: the question isn't whether AI is useful — it's who gets to set its rules.
- One paper argues that because a few firms own chips, cloud, data and apps together, they effectively set the rules for privacy, access and work — not governments.
- Real-world example: data centres in Ireland, the UK, the US and Chile bring heavy electricity and water use, but few local jobs and little local control.
- The fix proposed is bigger than antitrust: rights impact checks, forcing AI services to work with competitors, public research funding, and a real way for people to appeal decisions.
Why It Matters
Your AI choices, privacy and job conditions could soon be set by a few companies rather than elected governments.