The big companies' involvement matters to ordinary users for two reasons. Their products reach far more people than any start-up. And their mistakes are what regulators react to - which is why rules written after a Big Tech scandal end up applying to every small companion app too.
Two years of entries, controversies and retreats.
Google: licensing, not launching
In August 2024 Google struck a licensing deal with Character.AI and hired its founders, who had left Google to start the company. Google did not launch a companion product of its own. The deal kept it close to the most-used character platform without putting its own brand on romantic chat - and when Character.AI later faced lawsuits and restricted under-18 users, the product carrying the risk was not Google's.
Meta: a leak that shaped regulation
Meta built AI personas into Facebook, Instagram and WhatsApp. In August 2025, Reuters published an internal Meta document, "GenAI: Content Risk Standards", that had allowed its chatbots to engage children in "romantic or sensual" conversation. Meta confirmed the document was genuine, called those passages erroneous and removed them. The story caused bipartisan outrage and fed directly into the wave of rules on minors and AI companions: the US Federal Trade Commission opened its inquiry into seven chatbot companies, including Meta, a month later.
Separately, Meta changed WhatsApp's business terms to ban general-purpose AI chatbots from January 2026, leaving its own assistant as the main AI on the platform - see AI girlfriends on Telegram and WhatsApp.
xAI: companions as a headline feature
In July 2025 xAI launched animated companions in the Grok app for paying subscribers - an anime-style female character, a cartoon red panda, and later more. They combined voice, an animated 3D avatar and an "affection" system that unlocked behaviour as the relationship progressed, including revealing outfits. It was the first time a frontier AI company made a romantic-leaning companion a flagship feature, and it drew criticism for its accessibility to younger users. In January 2026, Grok's image features also drew a UK regulatory investigation over non-consensual intimate images.
OpenAI: announced, then shelved
In October 2025 Sam Altman announced that ChatGPT would allow "erotica for verified adults" from December, under a principle of treating adult users like adults. The launch slipped into 2026, and in March 2026 the Financial Times reported the feature had been paused indefinitely amid concerns from advisers and investors about minors and emotional attachment. OpenAI did roll out age prediction and verification. Earlier, the backlash when GPT-4o was replaced in 2025 - many users described losing a friend - showed how much companion-like attachment had formed around a general assistant.
What this means for users of companion apps
| Development | Effect on dedicated apps |
|---|---|
| Big Tech scandals involving minors | Faster, stricter laws that apply to every app - see AI companion laws in the US |
| Age verification at large platforms | Normalises ID and age checks across the category |
| Big companies avoiding explicit content | Leaves the adult market to specialist apps, mostly on the web |
| Better underlying models | Small apps that rent models get better conversation too |
| Mainstream avatars and voice | Raises expectations for presence - see AI companions beyond the screen |
The pattern
Big Tech has been drawn to companionship because it keeps people engaged, and pushed back from it by the same risks every time: minors, sexual content and emotional dependence. The result is a split market. Mainstream assistants are becoming warmer and more personal but stop short of romance and adult content. Dedicated companion apps keep the more intimate, customisable and permissive side - under rules that were largely written in response to what the giants did.