Under the AI Act, it is completely prohibited to use artificial intelligence to manipulate public opinion or create databases with biometric data such as facial features of people. For example, artificial intelligence used in critical infrastructure, banks, other financial institutions, or medical devices falls into the high-risk category. As explained by Lukas Benzul, director of the Czech Artificial Intelligence Association, companies and institutions falling into the high-risk category must ensure that AI is supervised by humans or tested regularly for its security.
“We need a high level of transparency and AI requires AI literacy among all employees using AI tools. It’s important to make sure people are aware that they are working with AI and understand the potential consequences.”
Jan Tomíšek, a lawyer and partner at Rowan Legal, recommends that large companies and institutions implement new roles in particular.
“While not required, we consider it practical to centralize this responsibility with someone who is familiar with all the key AI use cases within the organization. This person should be able to assess, categorize and determine where the responsibility lies in order to recognize the realms that it exists and exists.”
The rules make it easier for businesses that use AI only for everyday office work, for example. However, they still need to inform their clients that they are communicating with the AI. The regulations will be fully implemented by August 2026, and EU member states are now drafting their own laws to match them.
The Commerce Department will not add any regulations other than EU regulations to prevent the creation of more deficits. At the same time, the ministry is dealing with possible adjustments at the European level, Deputy Minister Jan Kavalilek said.
“What we’re worried about is that it shouldn’t be stricter because we can already see trends in the current draft that go beyond the recommended documentation. So far, the ministry has prepared a strategy for drafting the law, but the government has not yet approved it.”
The ministry would then like to discuss standards with corporate representatives and opposition parties so that it can be approved by the current Cabinet. However, there is a small chance that the law will reach lawmakers before the election. The new government will likely renegotiate it.
The ANO Party has no fundamental issues with the law. However, according to Vice-Chair Karel Havlichek, regulations in the field of artificial intelligence should be kept to a minimum.
“We need to understand that we are not a country of giants, but a country of small businesses, which are more comfortable, accessible and less intimidating if possible. We understand the need to be careful in the face of threats.
Zdeněk Zajicek, president of the Chamber of Commerce and Civic Democrat politicians, agrees. He said it would also need to be clear at the European level how authorities overseeing compliance with regulations work. In the Czech Republic, the Czech Communications Office is responsible for this. I think this is a good choice for Zazichek.
“The Czech telecommunications office wants to act in a predictable way, meaning that if they don’t follow the rules, they will have regular consultations so that businesses can prepare for what they are waiting for.
The Czech Communications Bureau should be able to test whether the artificial intelligence companies want to use meets the regulations. Havlíček is against the supervision of the Czech telecommunications office. In his opinion, the authorities are incompetent and the Ministry of Industry must take charge.
