AI can now expose the real creators of fake online accounts, threatening anonymity on the internet.

How Artificial Intelligence Uncovers Fake Accounts

In the network many users create anonymous or fake profiles, believing that this will protect their identity. However new research shows that modern AI models can not only detect such “forgeries,” but also identify the real owner.

Who studied this issue?

The Swiss Federal Institute of Technology in Zurich (ETH Zurich), the independent educational and research program MATS, and Anthropic jointly published the results as a PDF document.

How do AI models work?

The models use a range of signals: data matches (from small fragments to large datasets) and characteristic behavioral patterns. These cues allow them to “suspect” the authenticity of an account and, in most cases, find its real owner.

Key figures from the study

Metric Value
Successful fake detection 68 %
Accurate identification of the real owner 90 %

Why is this important?

The results can affect not only ordinary users but also those who publish information under other people’s names on social networks. AI can reveal the true identities of hackers who spread confidential employee data, as well as individuals asking questions on public forums. Even business leaders sometimes use pseudonyms to promote their ventures or harm competitors – a case known involving Elon Musk. Moreover, revealing identity can endanger whistleblowers, activists, and victims of violence for whom anonymity is vital.

Problems with protection against de‑anonymization

Scientists recommend several measures:

1. Restrict access to user data via APIs on platforms.
2. Strengthen mechanisms for detecting automated information gathering.
3. Minimize the export of large volumes of data.

However researchers note that even these steps do not guarantee 100 % protection. In the coming years – and even months – it will become increasingly difficult to keep AI from performing such operations.

Conclusion:

AI models are already capable of recognizing fake accounts and identifying their real owners in a significant number of cases. This threatens the anonymity of not only ordinary users but also those whose safety depends on secrecy. Strengthening protective measures may slow the process, but it will not halt technological development for now.