Google has sacked a senior software engineer after he claimed the company’s AI chatbot is a “sentient” being.
The search engine’s parent company Alphabet said Blake Lemoine’s claims about LaMDA (its language model for dialogue applications) are “wholly unfounded”.
“It’s regrettable that despite lengthy engagement on this topic, Blake still chose to persistently violate clear employment and data security policies that include the need to safeguard product information,” a Google spokesperson told Reuters.
Google’s LaMDA uses vast amounts of internet data to produce human-like responses to users’ demands.
The tech giant says it can “engage in a free-flowing way about a seemingly endless number of topics”.
Mr Lemoine, who worked at Google for seven years, began interacting with LaMDA as part of his job with the firm’s Responsible AI organisation.
He told The Washington Post he believed LaMDA was a self-aware person after testing whether it could use discriminatory or hate speech.
In April he shared a document with the company’s top executives entitled ‘Is LaMDA sentient?’ that included some of his conversations with it.
He was placed on administrative leave in June for violating Google’s confidentiality policy.
After news of his dismissal was first reported by the industry newsletter Big Technology, he has said he is now considering starting his own AI company to develop better storytelling in video games.