Connecticut: Bill SB 5 comes into force; AI companions are no longer allowed to pass themselves off as humans or seduce their users

Since 1 October, companion chatbots must identify risks of suicide, refrain from romantic relationships and excessive flattery, and AI subscriptions may no longer be renewed without written consent.

Signed on June 2 by Governor Ned Lamont, Connecticut's SB 5 law has been in force since October 1, 2026. It imposes a series of obligations on conversational companions: prohibition on pretending to be human, encouraging self-harm or violence, having a romantic or sexual relationship with the user, or showering them with praise and asking for gifts to maintain the connection. Services must detect signs of suicide risk, refer them to the 988 crisis line and refrain from offering psychological support themselves. They can also not be provided to minors without parental controls.

Subscriptions and whistleblowers

The text also targets paid offers such as ChatGPT or Google AI: their renewal now requires written notification and proof of the customer's consent. The breaches fall under the state's law on unfair commercial practices.

Finally, employees of border model developers will be able to anonymously report catastrophic risks, defined as those that could cause more than 50 deaths or more than a billion dollars in damage. Companies with more than $500 million in revenue have until January 1, 2027 to set up this channel.