HN.cz: Who Should Own AI? When a handful of private companies are building technology this powerful, we must set the rules before the damage is done.

Who should hold the reins over a technology that develops faster than any regulation can keep up?

This is a question I have been thinking about more and more lately, and one I find missing from Czech business conversations. While I read and hear these reflections increasingly from abroad, what I observe around me are topics about how to deploy AI, how to cut costs, replace people, gain an edge. We rarely have the debate about who is responsible for it and what rules should govern its use. Yet this is no longer just an economic question. It is an ethical and societal one too.

Previous breakthrough technologies of comparable force, nuclear weapons, GPS, the internet, mobile networks, emerged in the public sector or under direct government control. AI is the first exception. The most powerful models in the world are being built today by a handful of private companies with no mandate, no elections, no democratic oversight. And governments are arriving late.

I am not saying this is wrong. I am saying it is an unstable situation. And that we should be speaking openly about the risks and preparing for them before they catch up with us. Not as with social media, where we are addressing the negative effects on young people’s mental health, deepfakes and sexualised violence only at the last minute.

Regulation as Certification

The past year has brought several moments that exposed the potential risks of AI in full. Italy’s data protection authority temporarily blocked ChatGPT over the handling of personal data. The European Commission launched proceedings against several AI model providers under the new AI Act. The British Parliament opened a debate on algorithmic transparency. And in the United States, more than one congressional hearing has revolved around the question: who is responsible when AI fails?

I am not an advocate for regulation as a brake. As an entrepreneur and investor, that would be against the interests of business. However, I do advocate for regulation in the form of certification. Just as the state does not dictate what an airline should manufacture, but does require that a plane cannot take off without certification, mandatory audits of the most powerful AI models could exist before their deployment.

Another analogy is the certification of medicines and the requirement to demonstrate their safety before they are permitted to enter circulation. In the case of AI, mandatory transparency about the data on which models are trained should apply. And international rules that prevent regulation from becoming a tool of geopolitical advantage.

In Europe, we have a chance at this. The AI Act is a step in the right direction, even if an imperfect one.

When AI Helps Kill

The hardest part of this debate is military use. I cannot avoid it, because this is precisely where a clearly defined line of responsibility and ethics is critically important.

Worth attention is a story that has been circulating in recent months around Anthropic and their model Claude. The US military used, or attempted to use, Claude. Anthropic refused under conditions it considered unacceptable, and as a result found itself on a blacklist and faced public pressure, including verbal attacks from President Trump. Anthropic, however, held to its principle: the final decision is made by a human, not by AI. No autonomous weapons. No mass surveillance.

Stories like this show precisely why it is not enough to trust that companies have good intentions. What happens when models end up in the hands of undemocratic rulers or unethical companies?

In the Czech Republic, we barely discuss military use of AI. We leave it to the Americans, the Chinese, the Israelis, and then react. But if those of us who operate and invest in this technology cannot say where our own red line of responsibility lies, we have nothing to say about the societal one either.

Responsibility for the Impact on the Labour Market

The question of AI ownership and public debate about risks has one more dimension: less fatal, but far more widespread. Who bears responsibility for the impact on the labour market? If the most powerful models are being built and deployed by a handful of private companies, then the responsibility for their impact on our lives and the functioning of society also rests with them. Including the labour market.

Estimates vary, but none of them allow us to remain comfortable. Some analysts speak of the disappearance of tens of percent of entry-level white-collar positions within five years. Others point out that new jobs will emerge. Both can be true at the same time. The problem arises when the pie grows faster than people can adapt. Nobody is asking whose concern that is.

For me, one practical conclusion follows: the number one priority today is not to buy another AI tool. It is to educate the people who work here and to transform companies so that they can survive this wave. In my portfolio companies, I hold to one principle: AI is a new operational layer, not a new piece of software. A company that fails to understand this will pay twice: once for the licences, and once because the competition will have moved ahead.

Three Things I Would Like to Hear More Often

If I were to summarise what I, as an investor and entrepreneur, would like to hear more often in the Czech environment, it comes down to three things.

First: powerful technology cannot be held without oversight by either a handful of private companies or by the state alone. We need a sensible middle ground: mandatory audits of the most powerful models, transparent testing, international rules. In Europe, we can build something valuable on this, and we should be addressing that opportunity actively, not just waiting for the rules to come from elsewhere.

Second: we cannot escape responsibility. When I invest in a company that deploys AI, it is my obligation to ask and address where it can fail and who it will affect and how.

Third: we must talk about AI openly. Not only about the possibilities, but precisely also about the risks and the limits. If we leave the public conversation to technology enthusiasts on one side and alarmists on the other, we all lose.

Personally, I am an optimist and I see AI as an enormous opportunity for science, healthcare, education and the economy. However, optimism without responsibility is naivety, and that has no place in business or in the governance of public affairs.