As new “social players”, AI companions interact with people in ways that closely resemble human relationships, fostering strong emotional attachments. For the first time in human history, millions of people are forming emotionally meaningful relationships that are entirely controlled by corporations, which can freely alter, pause, or delete them. Such one-sided power may violate human rights standards protected by Article 8 of the European Convention on Human Rights (hereinafter the Convention), which guarantees the right to respect for private and family life. This should be considered by the European Court of Human Rights when developing a new practice to protect the Convention’s principles and standards.

In 2023, when Replika users lost their AI companions due to software updates, it was not a minor software issue for them, but the loss of a “wife,” a “best friend,” or “a better sexting partner”. A similar situation occurred in 2025, when the AI companion Dot shut down, prompting users to hold digital funerals, sharing final screenshots and farewell messages as if real people had died.

If an AI companion is becoming so psychologically significant to people, any interference with the human-AI relationship may affect the user’s emotional stability, mental well-being, or psychological integrity, thereby violating the rights protected by the Convention. One obvious example of such a violation is a violation of the rights declared in Article 8, which protects everyone’s private and family life and safeguards the right to establish and develop relationships with the outside world, as was mentioned by the European Court of Human Rights (‘the Court’) in Bensaid v. United Kingdom. Continuing this line of reasoning, the Court has held that mental health should be regarded as a crucial part of private life, associated with moral integrity, and protected under Article 8 of the Convention.

Due to lack of regulation, users of AI companions face uncontrolled third-party intervention in the human-AI relationship, potentially violating the principle of personal autonomy. As owners of AI companions, private companies can alter, suspend, delete, or redesign the human-AI relationship without any user’s consent and without legal consequences. Such uncontrolled company actions, in the absence of proper governmental control and regulation, can affect people’s moral integrity, psychological health, and the private life rights protected by the Convention.

Being obliged to protect human rights and freedoms enshrined in the Convention, countries have not only obligations to prevent government violations but also positive obligations to adopt measures to secure human rights, as recognised in cases such as X and Y v. The Netherlands. That is why the lack of governmental regulations governing human-AI relations, aimed at protecting the Convention’s standards and principles in such cases, also constitutes a human rights violation.

On the one hand, adopting instruments such as the Framework Convention and the EU AI Act gives rise to hope for human rights protection in Europe. But on the other hand, we should be realistic: we are only at the start of the way, and the one thing countries can do now is react quickly and sometimes be more flexible and less bureaucratic to address the challenges of these new times.

The lack of specific laws on AI companions in the EU, as well as the absence of a comprehensive set of regulations governing human-AI relations and protections, will likely prompt the Court and local courts in Europe to issue rulings without waiting for governmental regulations.

Very soon, we will face a growing number of the above-mentioned issues arising from the need to expand the scope of human rights protection in human-AI relations. Notably, while human-AI relationships may pose emotional risks and negative consequences for the person, they also play a very positive role, helping lonely and elderly people. That is why, in its early cases involving human-AI relations, the Court should be very careful in interpreting the Convention and in determining who is violating rights and freedoms and how. In some cases, the harm may be caused by AI companions, but in others by governments or by providers of AI companions.

The era of human-AI relations has only just begun and will bring many surprises. Human-AI relations are no longer solely a matter of software use or service provision but fall within the scope of rights and freedoms already protected by the Convention, so states should no longer avoid their obligations.