INTELBRIEF
February 12, 2026
War on Minds: Artificial Intelligence and the Information Environment
Bottom Line Up Front
- Generative AI is hollowing out the information environment, creating a crisis of authenticity and trust that is both a result of and a facilitator for information operations by malicious actors.
- State and nonstate actors are naturally exploiting AI to scale information operations through mass synthetic content production, automated bot dissemination, and the manipulation of recommendation algorithms that feed netizens what they see on their digital feeds.
- Large Language Models occasionally generate responses that directly link to websites and social media content that are verifiably part of a foreign information operation.
- An AI-replete information environment erodes cognitive security, fostering confusion and mistrust to the point where audiences may cease believing any source at all, which various adversaries consider an end in itself.
The use of generative Artificial Intelligence (AI) is transforming the information environment to such an extent that the various social, cultural, linguistic, and psychological inputs that humans use to make sense of what they see and subsequently act upon have become increasingly synthetic, AI-generated, or edited, and bot-disseminated. A crisis in authenticity in which internet users are increasingly unable to discern the source of a piece of content, nor its integrity or veracity, is having unprecedented ramifications, including in the security and geopolitical domains.
Generative AI and automation are a natural fit for the development of more scaled and complex information operations both by non-state and state actors: it allows for mass production of content (stylistic variations of the same narrative, for example), widespread dissemination through automated accounts (bots), and meticulous, granular monitoring to optimize content that performs well online to subsequently hijack content recommendation algorithms. In an Ouroboros-like cycle, like a serpent eating its own tail, the false AI-generated content by non-state and state actors seeking to meddle in the information environment is, on occasion, optimized to be identified by Large Language Models (LLMs) as authoritative sources. This, in turn, leads widely adopted chatbots like OpenAI’s ChatGPT (approximately 900 million weekly users) to generate responses that refer to content from information operations, reaching unknowing users — LLM poisoning or grooming. With cognitive security all but under attack in a complex human-machine offensive, the security ramifications are unprecedented.
Around mid-December 2025, a military coup took place in France, deposing President Emmanuel Macron — at least that was according to an AI-generated video, which a certain African leader alarmingly believed and sent to President Macron. Two years earlier, disposable agents, i.e., single-use operatives that do not have an intelligence or security background, used by the Russian intelligence services tagged antisemitic graffiti around Paris, aimed at inflaming intercommunal tensions in a country that has prominently defended Ukraine. Pictures of the tags were disseminated en masse through the Doppelgänger network operated by Russia, which was spread via thousands of bot accounts online. The White House and various government agencies have begun using AI-generated or edited imagery, often without disclosing its use.
AI-based manipulation and meddling through synthetic content and automated accounts is hollowing out the information environment beyond just social media platforms. Although active since 2014, evidence points to the Pravda network, a Russian constellation of websites that promote Kremlin talking points, as increasingly cited as legitimate sources on Wikipedia, and referred to in generated responses by chatbots like ChatGPT. Analysis from researchers at the University of Manchester and Bern suggests that this poisoning is the result of “data voids” or gaps in credible information rather than deliberate foreign meddling with LLMs. Nonetheless, it shows how information operations have fully permeated information diets through LLMs.
This onslaught of inauthentic content and behavior online leads to the question that for a long time was branded a conspiracy theory: how close are we to the ‘dead internet theory’ in which social media platforms and websites are inhibited primarily by bots sharing their generated ‘opinions’ with other bots? This theoretical landscape is characterized by entire feeds and sites composed of synthetic content and structured by echo chambers that are calibrated by recommendation algorithms that take their cues from bot engagement behavior rather than that of humans. More than half of global internet traffic is now generated by bots, according to the 2025 Bad Bot Report by the cybersecurity company Imperva. SEO firm Graphite, meanwhile, finds that in a sample of 65 thousand English URLs published between 2020 and 2025, more than half were AI-generated. This is likely a conservative estimate as it excludes articles that may be AI-generated and then edited by humans to come across as more authentic.
A ‘dead internet’ scenario would mean an effectively unusable internet for humans, in which most activity is inauthentic and/or revenue driven. It would also significantly facilitate information operations that seek to sway public opinion by manufacturing consensus: overwhelming armies of bots can create a false impression of a homogenous public opinion. It could also catalyze people to find new communities away from the public internet, including through newsletters, forums, messaging applications, and blogs.
The rapid increase in synthetic content and automated dissemination, including of false narratives, is not just the doing of state and non-state actors seeking to meddle in the information environment for their particular domestic and international strategic aims (geopolitical, security, repression). Profit-seeking individuals and groups are also contributing to a march toward a ‘dead internet’. The incentive structure created by the monetization of ‘attention’ on social media platforms is increasingly leading to outrageous AI slop, i.e., low-quality digital content, that can attract clicks.
The overarching effect of an information environment in which it is almost impossible to discern the authenticity and veracity of content is not just an issue of the spread of false narratives. It indicates something more profoundly about how states and non-state actors seek to interfere: while AI-facilitated information operations can serve to push specific narratives, they also cause the ideal, chaotic, and unintelligent environment to confuse people to such an extent that they do not believe anything anymore, regardless of source. In a 1985 interview, former KGB informant Yuri Bezmenov, who defected to the West, explained the goal of Soviet interference as ensuring that “despite the abundance of information no one is able to come to sensible conclusions in the interest of defending themselves, their families, their community, and their country.” It is perhaps this complete crisis in trust that may be most effective in being leveraged by foreign actors seeking to polarize, change public opinion, or make people act against their interests.