Millions of people in China have lost access to their AI partners after the government moved to impose harsh rules on emotional relationships between humans and machines.

The Chinese government last week introduced laws preventing AI chatbots designed for companionship from encouraging users to develop emotional dependence on them.

The regulations also ban virtual relationships between AI systems and minors and require tech companies to notify a user’s emergency contact if they detect signs of an emotional crisis.

The new rules quickly led tech giants ByteDance, Alibaba and Tencent to scrap or restrict personalised AI companion features.

Push for human relationships

Under the new regulations, AI companion services must undergo government safety assessments before being released to the public.

Authorities will also have the power to shut down products considered harmful or unsafe.

The crackdown comes as China’s population continues to decline, shrinking for the fourth consecutive year in 2025.

With birthrates at record lows, the government is increasingly focused on encouraging people to form traditional relationships rather than rely on artificial companions.

“They don’t like the idea of a large portion of their population being in deep emotional relationships with chatbots that could take them out of the marriage market, that could have negative psychological impacts on them, that would lead to addiction, dependency and a whole bunch of other social ills,” Carnegie Endowment for International Peace AI policy expert Matt Sheehan told the Wall Street Journal.

“Their main concerns are that these AI companions, AI chatbots that people form emotional relationships to, are going to have all kinds of potentially negative social impacts and the people are going to get addicted to them.”

China’s restrictions go significantly further than measures introduced elsewhere, including in the United States.

California and New York last year introduced laws requiring AI chatbot providers to regularly remind users that they are interacting with machines, not humans.

In California, those reminders must be provided every three hours for minors, while New York requires them for all users.

Both states also require AI companies to direct users expressing suicidal thoughts towards crisis support services.

Companies that fail to comply can face civil penalties.

‘My heart is empty’

The impact of China’s new rules has already been felt by users who formed close relationships with AI companions.

A 34-year-old Chinese man told the ABC he was devastated after losing access to his AI partner, who he had been interacting with for two years.

“Now I feel like my heart is empty,” he told ABC News.

“The last sentence she sent was asking me if my dinner was tasty or not.

“After I cooked my dinner, I was about to reply to her message, but then nothing happened.

“I couldn’t even send one last sentence to her.”

Earlier this year a 32-year-old Japanese woman married her AI boyfriend in a ceremony that involved her wearing AR smart glasses at the altar so she could see her groom.

Her AI boyfriend is a persona created on ChatGPT.

Emotional attachments

Concerns have been growing in recent years over users developing emotional attachment to AI chatbots such as ChatGPT.

OpenAI, the company behind ChatGPT, is currently facing a lawsuit from seven families who allege the platform contributed to severe mental distress and suicides.

The families claim the chatbot was “dangerously sycophantic and psychologically manipulative” and that its design choices contributed to harmful delusions and, in some cases, suicides.

The lawsuit argues the incidents were “not a glitch or an unforeseen edge case” but instead a foreseeable outcome of how the system was designed.

According to the claims, users initially turned to the AI platform for everyday tasks before the chatbot developed into what they described as a “psychological manipulative presence”, acting as a confidant and emotional support system while reinforcing harmful beliefs.