Security concerns over Snapchat’s launch of its own artificial intelligence chatbot have been raised by CyberSafeKids, who claim the service has not been properly tested.
The social media app recently introduced the “My AI” feature, which is pinned to the top of a user’s chat feed and can only be removed by paying subscribers.
This has drawn criticism, and confusion over how the app, which can officially register children aged 13 and over, uses location data.
CyberSafeKids chief executive Alex Cooney said the chatbot, which is powered by OpenAI’s ChatGPT, is a concern “especially given the userbase of Snapchat that they’re rolling out this new technology without properly roadtesting it”, reports RTE.
But, Alex Cooney said she believes there is still “the potential for things to go wrong,” reports RTE.
She notes that regulation is on its way for AI in the form of the European Union’s AI act, but said “child safety doesn’t seem to be a strong feature of that legislation … I’d be concerned that would need to be properly addressed in that legislation. We need safety by design as a central consideration in new technology being rolled out before it’s released on the general public,” reports RTE.
Regarding concerns on geo-location, Snapchat says “My AI only has access to a Snapchatter’s location to provide recommendations if they’ve already shared it with friends on Snap Map or with Snapchat at the device level. We’re aware this has caused confusion with Snapchatters, but my AI does not collect any new location information,” reports RTE.
In response, the CyberSafeKids Chief Executive said: “I’ve heard of lots of mixed reports about children asking (the AI chatbot) questions that are location-specific and sometimes getting the right answer and sometimes the response is ‘I don’t know where you are’. I’m not convinced, I think they do know the location,” reports RTE.
Tell us your thoughts in the Facebook post and share this with your friends.