OpenAI, the company behind generative artificial intelligence (AI) model ChatGPT, has said that it intends to launch a version of the chatbot that is safer for teenagers.
The move comes as regulators in many countries are starting to increase their scrutiny of chatbots.
Earlier this year, a 16-year-old Californian named Adam Raine died by suicide.
Raine’s parents have since filed a lawsuit against OpenAI, alleging that ChatGPT played a role in his death.
Writing on OpenAI’s blog, CEO Sam Altman said that the company needed to separate users who are under 18 and give them a different experience on ChatGPT.
The chatbot is intended for use by anyone aged 13 and over, but the new teen-friendly version will have stricter controls.
Altman gave examples, saying that ChatGPT would be trained not to participate in “flirtatious talk” with under-18s.
It would also avoid engaging in discussions involving suicide and self-harm, even in the context of creative writing.
Parental controls and age prediction technology will be used
Parental controls are expected to be introduced by the end of this month, which will allow guardians to link their accounts with a teen user (aged 13-17).
Open AI said that this will help guide how ChatGPT responds to their teen, based on “teen-specific model behaviour rules”.
Features such as memory and chat history can be disabled and “blackout hours” can be set, during which the teen will not be able to access ChatGPT.
Additionally, parents can receive a notification if the AI detects that the young person is in “acute distress”.
On rare occasions, if the parents cannot be reached, law enforcement may be involved.
As well as parental controls, the company is working towards an “age prediction” technology that will aim to estimate a user’s age by analysing how they use the system.
Acknowledging that such a system could be prone to mistakes, Altman said that ChatGPT would default to the under-18 experience if there was any doubt.
Adults would be given ways to prove their age, and in some countries, users might also be asked for ID.
Altman said that OpenAI was aware that its principles of freedom, privacy and safety were “in conflict” and that not everybody would agree with the move.
He added that they believed the “privacy compromise” for adults was “a worthy tradeoff” for safety.
Today’s news was brought to you by TD SYNNEX – the UK’s number one solutions distributor.
Destination AI from TD SYNNEX is our comprehensive end-to-end programme designed to help you and your customers get ahead of the AI curve.
Read more of our latest Artificial Intelligence stories
