Pintér Róbert, sociologist and information-society researcher and associate professor at Budapesti Corvinus Egyetem, argues that disinformation did not begin with generative artificial intelligence (AI), although AI provides new tools for producing manipulated content. He has studied the social effects of digitalization for more than twenty-five years, recently focusing on AI, corporate digital transformation and disinformation.
Where the problem begins
Pintér emphasizes that one must look beyond technology to the social media environment where algorithms, advertising systems and economic incentives shape how information spreads. Because many platforms are ad-driven, user attention becomes a commodity and algorithms optimize for engagement — which can encourage disinformation production.
He notes that motivations are mixed: political, economic, wartime and other information-driven aims all play a role. As an illustrative case, during the 2016 US election campaign Macedonian youths earned considerable revenue from fake news sites without necessarily intending to influence the election politically, yet their actions had political effects.
Treating disinformation as a chain
In his study, Pintér describes disinformation as a chain with multiple intervention points from motivation through creation to distribution. He argues there is no single silver-bullet solution and no sole actor responsible for fixing it.
- Motivation: reduce economic incentives so that traffic built on falsehoods is not easily monetized.
- Creation: labeling AI-generated content is one intervention, and European regulation attempts to address this; however, these measures can be circumvented (for example by running open-source models locally or by actors with malicious intent).
- Distribution: Pintér considers this the most important stage. In traditional media, editorial responsibility and fact-checking matter; on platforms, algorithmic and advertising logics must be constrained because services can profit even when disinformation spreads.
Fact-checking is necessary but not sufficient: people often assimilate information into their worldview, and repeated exposure can have an effect even when the information is known to be false.
Deepfakes, economic fraud and trust
Pintér warns that AI enables new forms of abuse with significant economic consequences. He cites a 2024 case in Hong Kong in which an employee of a multinational company was reportedly persuaded by a deepfake video call to transfer roughly 25 million US dollars to fraudsters.
He also points out that similar schemes existed well before generative AI: in the early 2010s a company nearly fell victim to a transfer requested in the CEO’s name, and the fraud was discovered only when someone questioned the suspicious amount.
One of the biggest societal risks of deepfakes is erosion of trust: if anything can be plausibly labeled fake, genuine evidence can be dismissed as manipulated. Trust, Pintér notes, is built very slowly and lost quickly — an attack, ransomware incident or poor communication can rapidly damage an organization’s credibility.
Education, research and the "Brain First" approach
Pintér argues that AI can be a useful tool but thinking should not be outsourced to it. He recommends a "Brain First" approach: first think independently and draft your own outline, then use AI as an assistant to identify blind spots, coherence issues or missing elements.
In teaching, he gives students tasks to prepare summaries without AI first, then refine them together with AI. This typically increases the amount of work and deepens engagement rather than reducing effort.
In research, AI can speed literature search, compile bibliographies or edit texts, but researchers must verify AI-generated claims and check original sources. If a paper cites a statistic (for example, "82 percent"), researchers must find that figure in the original source to avoid relying on hallucinated outputs.
From techno-optimism and techno-pessimism to digital realism
Pintér observes two common reactions to digitalization: techno-optimism, which expects technology to solve social problems, and techno-pessimism, which foresees catastrophe. He advocates "digital realism": technology is neutral in itself, and social, economic and political contexts determine its effects.
The early internet myth that it would self-regulate without oversight has largely been replaced by market dynamics and profit motives, bringing new problems rather than automatically improving society.
Conclusions
Pintér concludes that while AI provides new means for creating and distributing disinformation, responses must be multi-layered: regulation, platform accountability, changes in economic incentives, media transparency, education and stronger critical thinking are all necessary. The goal should be to reduce the number of people exposed to the harms of disinformation and to limit societal damage.
The article’s publication was supported by Budapesti Corvinus Egyetem.
