As EU moves to regulate artificial intelligence, poll finds half of Czechs AI-skeptic

The EU is going to regulate the use of artificial intelligence so it does not violate people’s civic freedoms.

Marcel Kolaja, a Czech member of the European Parliament, said the use of artificial intelligence needs more attention from politicians since the development is very fast and the impact on society will be large in the coming years. Kolaja is a software engineer and deputy chairman of the Pirates.

The development of AI makes 48 percent of Czechs uneasy while about a quarter are "excited and worried in equal measure," according to a recent survey carried out by Charles University’s Faculty of Social Sciences and the polling agency Ipsos.

Kolaja said most uses of AI are unproblematic from an EU perspective, but there are cases that the European Commission considers risky such as facial recognition technology used by law enforcement agencies or e-proctoring, a tool that schools can use to check that students are not cheating when taking exams remotely.

E-proctoring monitors the camera and microphone of a student's computer to assess whether the student is texting, talking to someone, or holding a mobile phone. "There's a danger that the system will malfunction," Kolaja said, according to the Czech New Agency (ČTK). He has pushed for e-proctoring to be considered high-risk in the planned regulation because of the risk of discrimination against people with disabilities or darker-than-average skin tones.

One of the most controversial areas of AI use is facial recognition technology by law enforcement authorities to identify suspects. "The EC has even proposed a ban with some exceptions," Kolaja said.

One rule being considered would make evaluating camera footage possible with a court order only. The EU considers the practice of "social credit," known from China, where a facial recognition system is linked to scoring people's behavior, to be completely unacceptable, he said.

Kolaja said that AI regulation could be approved by the end of the year. He noted that member states differ in their approach to AI instruments and the potential problems associated with them. Italy, for example, recently banned the popular chatbot ChatGPT on its territory due to privacy concerns.

The survey by Charles University and Ipsos showed that six out of 10 people would support a call signed by, among others, U.S. entrepreneur Elon Musk, which called for a six-month suspension of the development of generative AI such as the ChatGPT model.

The survey focused mainly on artificial intelligence in the media. About 40 percent of the Czech population said they come into contact with AI at least once a day, with people most often correctly identifying that they encounter AI in products related to human health such as fitness trackers or when shopping online.

But most respondents did not know that AI could write news articles in Czech and that some media outlets were already taking advantage of this.

ČTK, for example, uses automatic text generation to create simple texts for news about gasoline prices or election results. About half of the people were able to correctly identify that a report on the price of gasoline was generated automatically.

In the case of radio news, an experimental audio recording of a Czech Radio robotic presenter reporting on traffic news was misidentified as human by 54 percent of Czechs. Another group of respondents heard the voice of a real presenter. In this case, 9 percent of them thought it was a robot.

Czechs are rather reticent in actively testing artificial intelligence. The freely available ChatGPT system was tried by only 15 percent of the population. People under the age of 24 have the most personal experience, with about half trying this technology. Men and people with higher education also showed greater interest.