Google Fires Researcher Who Claimed LaMDA AI Achieved Sentience

Blake Lemoine Fired from Google Following Claims of AI Sentience

Blake Lemoine, an engineer with seven years at Google, has been terminated from his position, as reported by Alex Kantrowitz in the Big Technology newsletter. Lemoine allegedly revealed his firing during a recording of his podcast, though the episode has yet to be released. Google has confirmed his dismissal to Tech media.

Lemoine was most recently involved in Google’s Responsible AI initiative and made headlines last month by asserting that the company's AI project, LaMDA (Language Model for Dialogue Applications), had achieved sentience. LaMDA was launched by Google last year to enhance computer-generated open-ended conversations. Lemoine not only asserted that LaMDA had acquired sentience but also questioned whether it possessed a soul. He reportedly stated to Wired, "I legitimately believe that LaMDA is a person."

After making these controversial claims to the media without authorization, Lemoine was placed on paid administrative leave. Google has consistently denied that its AI is sentient, supporting this stance in communications with the Washington Post.

The AI research community reacted strongly against Lemoine's assertions. Margaret Mitchell, a former Google employee dismissed for advocating diversity, tweeted that systems like LaMDA do not develop intent; rather, they model human communicative intent through text strings. Gary Marcus dismissed Lemoine's statements as "nonsense on stilts."

In response to Lemoine’s claims, Google issued a statement to Tech media, emphasizing its commitment to responsible AI development. The company noted that LaMDA underwent 11 reviews and detailed its responsible development in a published research paper. Google expressed regret over Lemoine's failure to adhere to clear employment and data security policies and reiterated its commitment to carefully advancing language models while wishing him well in the future.