Troll farms are a thing of the past. Now we are threatened by AI content farms. Experts are sounding the alarm

Websites with false content generated by AI may be destructive to society, negatively affecting trust in the media and the recognition of disinformation – told PAP prof. Aleksandra Przegalińska and the president of ZPWC, Maciej Kossowski.

Prof. Przegalińska a victim of AI. It's an alternative reality that has a life of its own

In December 2025, specialist in the field of artificial intelligence, prof. Przegalińska informed on website X about the existence of a number of websites run by one entity, which contain false content about her, mass-generated using AI. According to the expert, these “content farms” also publish fake news about other public figures, including: actors and politicians, including Deputy Prime Minister and Minister of Digital Affairs Krzysztof Gawkowski, and about well-known Polish and foreign brands.

“After my intervention, specific entries disappeared, but the websites themselves continue to operate because their logic does not concern one person or one text. These are content factories that immediately produce new materials in place of the deleted materials. For their creators, a single correction or report is only an operational cost, not a real barrier,” Prof. told PAP. Przegalińska.

She noted that when she saw false content about herself, she was surprised, irritated and felt helpless. “This experience is disturbing because it makes you realize how easy it is to create an alternative version of reality that begins to take on a life of its own, regardless of the facts and intentions of the people concerned,” she emphasized.

Cheap, fast and destructive. How artificial intelligence destroys trust in the media

“Such websites cannot be considered journalism under any circumstances,” said Maciej Kossowski, president of the Association of Digital Publishers Employers (ZPWC), which brings together the largest media companies in Poland. As Kossowski emphasized, websites with low-quality content have been created for years. Launching them was not difficult, but people were needed for it – he pointed out.

“Now, with AI, it is trivial. One person can create dozens of content and websites on any topic,” he noted.

According to prof. Przegalińska, this type of portals are created mainly for economic and political reasons.

“Generating mass content through AI is cheap, fast, and easy to monetize through advertising or traffic manipulation,” she noted.

She estimated that the people behind such websites often want to influence the public debate, test the narrative or cause information chaos, favoring those who can manage it.

“The consequences of the existence of such websites are serious,” she emphasized. As she said, for individuals this means “reputation risk, stress and the need to constantly respond to false information.” However, for society, this means – according to the expert – “erosion of trust in the media, experts and institutions, and in the long run also a weakening of the ability to distinguish knowledge from disinformation.”

“How dangerous modern tools can be in the hands of foreign services”

According to the president of ZPWC, the case of prof. Przegalińska should be treated as a warning. “These are primitive activities aimed at making money, but they should make us realize how dangerous modern tools can be in the hands of foreign services that want to destabilize societies,” Kossowski emphasized. “There are many examples. Let's imagine a 'fake' website with medical or psychological advice that no one verifies and what consequences such activity may have,” he said.

According to experts, combating this phenomenon requires several parallel actions, including legislative ones. “We need wise regulations that will not kill innovation, but will limit the spread of technologies that may be destructive to societies,” Kossowski said.

As noted by lawyer and expert in the field of AI implementation, Dr. Michał Nowakowski, the legal environment for the legal and ethical use of artificial intelligence “is at an infant stage.” In his opinion, the AI ​​Act, the EU regulation on artificial intelligence, “helps little” in such situations. The act includes, among others: regulations on prohibited practices, transparency obligations, including marking AI-generated content, but – according to the expert – it is not known whether entities will comply with these regulations at all after they enter into force.

Dr. Nowakowski assessed that the mere adoption of an act on artificial intelligence in the EU is “definitely not enough”, because the problems associated with the use of generative AI go “far beyond the scope of this regulation”. “Today's challenges also involve, among others, creating deepfakes, violating intellectual property rights, and unauthorized use of image or voice,” he said.

The first lawsuit for the voice of AI in Poland. Lector versus septic tank manufacturer

In the latter context, the lawsuit of lecturer Jarosław Łukomski against the company JFC Polska, which produces septic tanks and sewage treatment plants, may be considered. Łukomski accuses the company of processing his voice using AI and then using it in advertising. The company – according to media reports – claims that it has not committed any illegal activities or violations. This is the first lawsuit of this type in Poland.

In addition to the AI ​​Act, which has not yet been fully implemented in Poland, people injured in cases related to generative AI are protected by civil law, in the part regarding personal rights – explained Dr. Nowakowski. Personal rights include, among others: health, freedom, honor, freedom of conscience, name or nickname and image. As the lawyer pointed out, violating personal rights may also constitute a crime, e.g. defamation or insult.

“In such a situation, in addition to civil instruments, we are entitled to solutions of a criminal nature – above all, we can and should submit an appropriate notification about the suspicion of committing a crime,” said Dr. Nowakowski. He added that support can also be sought from online platforms that “should strive, for example, to remove harmful advertisements, especially since they are obliged to act by EU regulations.”

As the expert pointed out, in some cases of violations related to AI, press law will also apply, but it depends on whether we are dealing with a publisher. “And this may be a problem in the context of unspecified publishers who do not meet high standards, i.e. good practices and ethical norms, and do not respect legal regulations” – emphasized Dr. Nowakowski.

We need a strong state that will be able to enforce regulations

“Even if we create very good regulations – there are such attempts – their effectiveness will be determined by the effectiveness of courts or law enforcement bodies. In a world where various companies and people operate from, for example, China or from jurisdictions with a weaker legal regime, it will be difficult to enforce these laws,” he said. In his opinion, Internet, domain and online platform providers will play an important role in this process.

President Kossowski also drew attention to the importance of platforms. In his opinion, large technological platforms, such as search engines or social media, must be responsible for the content published on them. “They can no longer act as intermediaries exempt from liability. They play too important a role in the information ecosystem,” he emphasized. He noted that a “strong state” is needed that will be able to enforce regulations.

Kossowski pointed out that business self-regulation is also necessary. “I mean companies such as OpenAI or Google, which should adopt very high security standards,” he noted. He recalled that OpenAI has introduced restrictions in ChatGPT, as a result of which AI no longer generates individualized medical advice, which – according to the president – is a “good direction”.

According to Kossowski, public education and “support of credible media” are also needed. As he explained, it is not about financial support, but about the state creating market and legislative conditions that will allow them to function. “We must be aware that 'fake' websites, together with the platforms, take part of the advertising money at very low costs. On the other side, there are the reliable media that spend a lot of money on editorial offices, information verification, content publication, etc.” – he emphasized.

Prof. Przegalińska, in turn, noted that “legislation is necessary, but it is not enough because the law always reacts slower than technology.” In her opinion, the following are also key: platform responsibility, transparency of content sources, media education and developing critical thinking skills.

“Without this, even the best regulations will only be a partial solution,” the expert emphasized.

The Artificial Intelligence Act (AI Act) entered into force on August 1, 2024, with certain exceptions. In February 2025, they came into force, among others: regulations on AI systems that are banned in the EU. Member countries had until August 2 last year to establish a market surveillance authority. August 2 this year almost the entire act will apply (except for one article), including provisions on the correct labeling of AI-generated content. The full act will come into force in 2027.

The draft Act on Artificial Intelligence Systems, prepared by the Ministry of Digitization, which is to implement the AI ​​Act, was published for the first time on October 16, 2024. Work on it is still ongoing – in mid-December last year, the Ministry published two versions of the draft and sent them to the Standing Committee of the Council of Ministers.

Monika Blandyna Lewkowicz (PAP)

mbl/ mrr/