Hackers and scammers are tired of AI spam: their private forums are flooded with AI slop

Cybercriminals have faced an unexpected problem: even closed forums and the darknet have been flooded with AI-generated content. This has forced hackers and scammers to complain about the proliferation of neural networks in places where they try to discuss cyberattacks and illegal activity.

When spam reaches the underground

AI-generated content is now everywhere. Neural networks generate articles, comments, posts, and forum messages so quickly and cheaply that traditional platforms like Reddit, Twitter, and YouTube have long been oversaturated with this material. But now the wave has flooded specialized closed forums, where cybercriminals gather to discuss their activities.

Closed communities of cybercriminals were long considered relatively safe havens. There are fewer government special services, better moderation (among their own), higher average quality of information. Members of such forums discussed real hacking methods, sold stolen data, shared best practices, knowing that their content would not reach outsiders. But AI spam has reached there too.

How low prices created a problem

• Cost of text generation via API has fallen to the bottom
• Spambots can be programmed in hours using ready-made models
• Low barriers to entry for malicious actors who want to flood a forum
• Lack of specialized protection against AI automation on closed platforms
• Rapid growth in accessibility of powerful language models through APIs

The problem is that neural network models are now cheap enough and simple to use. A spambot can be assembled in a few hours using the OpenAI API, Claude API, or freely available models on Hugging Face. The cost of generating 10,000 forum posts is now pocket change. Spambots are flooding forums with generated content—from meaningless posts to boost activity counters to more sophisticated schemes where an AI bot mimics an experienced member to build reputation and then sell a compromised account.

At the moment, this is even cheaper and more effective than hiring people to perform the same task.

The paradox of technology

There is a certain irony in this situation. The community of cybercriminals, which actively uses technology and automation for its own purposes, now suffers from the consequences of the development of completely different technologies. Hackers were among the first adopters of automation, bots, scripts, and AI solutions. But nobody expected this wave to turn against them.

Historically, cybercriminal forums were several steps ahead of the mainstream internet in terms of technological solutions. They implemented encryption, anonymity, and decentralization before these concepts became popular. But AI spam is a problem that scales much faster than any traditional moderation can handle.

"Even in the underground internet, there is no more peace," forum

members are complaining along these lines.

What this means

The spread of AI-generated content has already reached a scale where it affects even the niches that are trying their hardest to hide from it. This is a telling signal that the problem of information quality and content control is no longer a separate task of major technology platforms, but a global reality that affects literally everyone.

Cybercriminals are forced to seek new ways to protect themselves from AI automation, just as regular users and organizations do. The irony is that they can use the same tools as their enemies: more advanced AI detection models, new user verification methods, decentralized architectures. But this is a problem that they clearly did not expect to solve.