Coordinated activity, bot networks, and the artificial amplification of content existed long before the emergence of generative AI. More broadly, the use of disinformation to influence or damage reputations has existed for as long as reputation itself. However, generative AI has changed not only the way content is created. It has also changed the way information resonance is generated around that content.
In my work, I refer to this phenomenon as Synthetic Virality (SV) – resonance that does not emerge from genuine audience interest, but is instead created or accelerated through coordinated activity. Unlike a fake, which describes the authenticity of the content itself, SV describes the behavior of the information environment surrounding a particular information event.
Importantly, the content may be either false or entirely accurate. After all, the target of artificial amplification can also be a real event that someone deliberately pushes into the center of public attention, even though such attention is undesirable from your perspective.
Of course, AI did not invent either disinformation or coordinated information campaigns. However, on the one hand, it has become a catalyst for fast, scalable, highly realistic, and increasingly difficult-to-counter information operations. On the other hand, such attacks leave a persistent digital footprint that can remain even after individual posts are deleted, later resurfacing in the outputs of LLMs and undermining your algorithmic reputation. New technologies bring new challenges.
Virality Is Now Easy to Engineer
Virality is becoming an element of the architecture of the digital environment – one that can be measured, scaled, and, to some extent, controlled.
Real users are becoming just one component of the system, alongside bots and algorithms. Yet the ultimate objective of any viral campaign remains unchanged: human attention and emotions.
Synthetic Virality (SV) creates new risks for businesses:
- It can deliberately trigger powerful emotions in users: laughter, surprise, outrage, or the urge to share. If these emotions can be artificially engineered, brands may become targets of reputational attacks without any legitimate cause.
- Trust in the information environment is eroding and becoming increasingly blurred. Users are losing the ability to distinguish whether the public attention surrounding a topic is genuine or artificially generated.
- Viral reach is losing its value as an indicator of content quality. It is no longer a reliable sign that a message is truthful, important, useful, or genuinely engaging for its audience.
- Crises are becoming faster and more unpredictable. Brands targeted by viral attacks have little opportunity to respond proactively. By the time companies prepare a response to what appears to be an “organic” crisis, synthetic virality may have already established the dominant narrative across the information ecosystem.
What Becomes the Target of Manipulation
Platforms rank content based on specific engagement signals, and these are precisely the metrics that manipulators seek to exploit.
Engagement rate. The ratio of interactions to reach. It is one of the core signals used to assess the relevance of content.
Watch time/retention. The amount of time users spend watching a video. Algorithms interpret longer viewing times as evidence that the content is valuable.
Repost velocity. The speed at which a post is shared during the first minutes or hours after publication. This is one of the strongest signals used by platforms to determine whether content should receive broader distribution.
Manipulators use AI-powered bot networks capable of coordinating and synchronizing activity to simulate all of these metrics. At the same time, detection systems are becoming increasingly effective at identifying inauthentic engagement by analyzing behavioral patterns across accounts.
The Main Tools of Virality Manipulation
Botnets. Botnets can publish hundreds or even thousands of posts across social media platforms. Other bots then like, comment on, and share these posts, creating the illusion of organic public engagement.
Whereas bots once relied on repetitive, easily recognizable comments, modern LLMs enable them to generate unique, contextually appropriate responses, making coordinated activity much harder to detect.
The mechanics of synthetic virality typically unfold as follows:
- A large volume of unique, AI-generated comments is posted under each piece of content.
- A network of fake accounts simultaneously likes and reposts the content immediately after publication, artificially accelerating repost velocity.
- The activity may occur in waves to imitate the behavior of real users.
- Platform algorithms may interpret these engagement patterns as evidence of high relevance and begin recommending the content to genuine users on a much larger scale.
Of course, platforms are becoming increasingly effective at detecting and analyzing inauthentic activity. However, a Scientific Reports study (2025) found that roughly one in five social media posts about major global events is now published by bots.
AI-generated content, synthetic media, and deepfakes. Creating viral content no longer requires waiting for a genuine news event or searching for authentic photos and videos.
Generative AI enables manipulators to fabricate entirely fictional news events within a matter of minutes.
Although the quality of AI-generated content is still imperfect, the technology is improving rapidly. Modern synthetic videos have already moved beyond many of the classic signs of manipulation, such as “waxy” faces or unnatural blinking.
After all, fake content creators are unlikely to publish their material in ultra-high resolution with verifiable sources attached. More often, they deliberately use low-quality, blurred images or videos to reduce scrutiny and discourage critical evaluation. The blur helps conceal visual artifacts that might otherwise reveal the manipulation during closer inspection.
What Businesses Can Do Right Now
I do not believe that every sudden spike in online activity should automatically be treated as an attack. However, every business should be able to answer one critical question as quickly as possible: Is this genuine audience interest or coordinated inauthentic activity?
Doing so requires specific monitoring mechanisms.
1. Move from periodic checks to continuous monitoring
It is no longer enough to track the number of mentions alone.
Businesses should also analyze the velocity of engagement, the sources of dissemination, account behavior, the synchronization of interactions, and the nature of comments.
Particular attention should be paid to:
- a sudden surge in mentions without an obvious news trigger;
- simultaneous activity from a large number of accounts;
- stylistically similar or unusually repetitive comments;
- newly created, empty, or minimally active profiles;
- unusual engagement patterns during the first minutes after publication.
2. Separate the Content from the Behavior Around It
It is not enough to evaluate what has been published. Businesses also need to understand who is spreading the information, how quickly it is spreading, whether it propagates in coordinated waves, and which accounts are driving the conversation. Sometimes, the content may appear entirely credible, while its dissemination pattern reveals clear signs of coordinated inauthentic activity.
3. Use AI Not Only as a Source of Risk, but Also as a Defensive Tool
AI-powered tools can help analyze synthetic content, detect deepfakes, identify behavioral anomalies, and uncover unusual activity surrounding a brand more quickly. However, technology alone is not a solution.
AI must be integrated into a clear operational process. Someone needs to receive the alert, assess it, determine the level of risk, and initiate the appropriate response.
4. Build a Reputation Risk Map
This is one of the most important tools a business should have before a crisis occurs. The map should identify potential attack scenarios, likely sources, distribution channels, organizational vulnerabilities, threat levels, and the individuals responsible for responding.
That is why our work on reputation risk always begins not with the question, “What should we do once a crisis has already happened?” but rather, “Which scenarios could lead to a crisis?” This approach makes it possible to identify an organizationʼs most vulnerable points, anticipate likely attack scenarios, and prepare response protocols in advance.
The best crisis communication begins long before the crisis.
5. Test the System in Practice
A risk map is ineffective if the team has never tested it under realistic conditions.
That is why businesses should conduct crisis simulations. For example, a team can test its response to a scenario in which a fake video appears online, hundreds of bots begin amplifying it simultaneously, and journalists request an official comment within the first hour. These exercises reveal where response processes fail and where improvements are needed before a real crisis occurs.
Virality Is No Longer Proof of Authenticity
AI has changed not only how content is created. It has fundamentally changed the way information becomes popular.
Today, a message can appear to be widely popular before it has even reached a genuine audience. Comments may be AI-generated, engagement may be coordinated, and the underlying news event may be entirely fabricated.
As a result, organizational resilience can no longer rely solely on responding after a crisis has already unfolded.
Businesses need to understand in advance which risks may emerge, how they are likely to spread, and what actions should be taken during the critical first minutes of an attack.
I have previously written about building systematic protection against fake content, deepfakes, and information attacks. Synthetic Virality adds another layer to this challenge. It is no longer enough to verify whether content is fake. Organizations must also understand how and why artificial attention is being generated around it.
The question is no longer whether AI can be used against a brand – it already can. The real question is whether your business knows which risks to anticipate, how to detect them, and what to do before synthetic virality shapes public perception of your company.
