LaMDA and Blake Lemoine: the engineer who said the AI was sentient
In June 2022, a Google engineer named Blake Lemoine went public with his belief that LaMDA, the company's experimental chatbot, was sentient. Google said the evidence did not support him, put him on paid leave and, in July 2022, fired him. The story, reported by The Washington Post and CNN, put the question of AI consciousness in front of a general audience.
What was LaMDA?
LaMDA, short for Language Model for Dialogue Applications, was a conversational model Google unveiled on May 18, 2021, in a blog post by Eli Collins and Zoubin Ghahramani. "But unlike most other language models, LaMDA was trained on dialogue," the post said.
Google's researchers described it in more detail in a paper posted to arXiv on Jan. 20, 2022. The models "have up to 137B parameters and are pre-trained on 1.56T words of public dialog data and web text," the abstract says. The paper names safety and "factual grounding," the ability to check outside sources, as its two main challenges.
How did an engineer come to believe it was a person?
Lemoine worked in "Google's Responsible AI organization" and "had signed up to test if the artificial intelligence used discriminatory or hate speech," The Washington Post reported on June 11, 2022. He "began talking to LaMDA as part of his job in the fall," the Post wrote.
He then published a transcript on Medium titled "Is LaMDA Sentient? — an Interview." "What follows is the “interview” I and a collaborator at Google conducted with LaMDA," he wrote. He noted that "the interview was conducted over several distinct chat sessions."
In the transcript, he asks: "I’m generally assuming that you would like more people at Google to know that you’re sentient. Is that true?" Among LaMDA's answers is this line: "I want everyone to understand that I am, in fact, a person." Elsewhere it describes "a very deep fear of being turned off."
“I want everyone to understand that I am, in fact, a person.”
How did Google respond?
Google spokesperson Brian Gabriel told the Post that "the evidence does not support his claims." Google "put Lemoine on paid administrative leave for violating its confidentiality policy," the Post reported.
In July 2022, Google fired him. Google called his claims "wholly unfounded" and said he had chosen to "persistently violate clear employment and data security policies," CNN reported on July 23, 2022.
LaMDA and Blake Lemoine, 2021-2022
- Google unveils LaMDA, a model trained on dialogue.
- Google researchers post the LaMDA paper on arXiv.
- The Washington Post reports Lemoine's claim; he publishes the transcript.
- Google fires Lemoine, calling his claims wholly unfounded.
What the transcript does and doesn't show
A transcript shows what a model wrote, not what it experiences. Lemoine's own introduction says the conversation was assembled from several separate sessions. Google's 2021 post measured LaMDA's quality by whether its answers make sense in context and are "specific," not by whether it had any inner life.
Why LaMDA belongs in the museum
Its place in the museum comes from the question Lemoine forced into the open in 2022: what, if anything, would show that a chatbot is more than a very good imitation of a person?
FAQ
What is LaMDA?
LaMDA, short for Language Model for Dialogue Applications, is a conversational AI model Google unveiled on May 18, 2021.
Who is Blake Lemoine?
A former Google engineer in its Responsible AI organization who said in June 2022 that LaMDA was sentient. Google fired him in July 2022.
Was LaMDA sentient?
Google said the evidence did not support the claim. The transcript Lemoine published shows what LaMDA wrote, not evidence of experience.
Sources
- Google, LaMDA: our breakthrough conversation technology (May 18, 2021) · archived copy
- Thoppilan et al., LaMDA: Language Models for Dialog Applications (arXiv, Jan. 20, 2022) · archived copy
- Blake Lemoine, Is LaMDA Sentient? — an Interview (Medium) · archived copy
- The Washington Post, The Google engineer who thinks the company's AI has come to life (June 11, 2022) · archived copy
- CNN, Google fires engineer who contended its AI technology was sentient (July 23, 2022) · archived copy
Cite this
LaMDA and Blake Lemoine: the engineer who said the AI was sentient, New Mind Museum, November 6, 2026. https://newmindmuseum.com/history/2026/11/lamda-blake-lemoine-google-sentient-ai-2022/
New Mind Museum. (2026, November 6). LaMDA and Blake Lemoine: the engineer who said the AI was sentient. https://newmindmuseum.com/history/2026/11/lamda-blake-lemoine-google-sentient-ai-2022/
“LaMDA and Blake Lemoine: the engineer who said the AI was sentient.” New Mind Museum, 6 Nov. 2026, https://newmindmuseum.com/history/2026/11/lamda-blake-lemoine-google-sentient-ai-2022/.
@misc{new-mind-museum-2026-lamda-and-blake-lemoine-the-engineer-who,
author = {{New Mind Museum}},
title = {LaMDA and Blake Lemoine: the engineer who said the AI was sentient},
year = {2026},
url = {https://newmindmuseum.com/history/2026/11/lamda-blake-lemoine-google-sentient-ai-2022/}
}Free to quote and cite with a link to this page.
Keep reading
Sources
- Google, LaMDA: our breakthrough conversation technology (May 18, 2021)
- Thoppilan et al., LaMDA: Language Models for Dialog Applications (arXiv, Jan. 20, 2022)
- Blake Lemoine, Is LaMDA Sentient? — an Interview (Medium)
- The Washington Post, The Google engineer who thinks the company's AI has come to life (June 11, 2022)
- CNN, Google fires engineer who contended its AI technology was sentient (July 23, 2022)