LaMDA, which stands for Language Model for Dialogue Applications, is a family of conversational neural language models developed by Google. The first generation was announced during the 2021 Google I/O keynote, while the second generation was announced at the following year's event. In June 2022, LaMDA gained widespread attention when Google engineer Blake Lemoine made claims that the chatbot had become sentient. The scientific community has largely rejected Lemoine's claims, though it has led to conversations about the efficacy of the Turing test, which measures whether a computer can pass for a human.

History

Announcements

Google announced the LaMDA conversational neural language model during the Google I/O keynote on May 18, 2021, powered by artificial intelligence.^[1] Built on the Transformer neural network architecture developed by Google Research in 2017, LaMDA was trained on human dialogue and stories, allowing it to engage in open-ended conversations.^[2] Google states that responses generated by LaMDA have been ensured to be "sensible, interesting, and specific to the context".^[3]

On May 11, 2022, Google unveiled LaMDA 2, which serves as the successor to LaMDA, during the 2022 Google I/O keynote. The new incarnation of the model draws examples of text from numerous sources, using it to formulate unique "natural conversations" on topics that it may not have been trained to respond to.^[4] Additionally, Google launched the AI Test Kitchen, a mobile application powered by LaMDA 2 capable of providing lists of suggestions on-demand based on a complex goal.^[5][6] Originally open only to Google employees, the app was set to be made available to "select academics, researchers, and policymakers" by invitation sometime in the year.^[7] In August 2022, the company began allowing users in the U.S. to sign up for early access.^[8]

Sentience claims

Lemoine's claims that LaMDA may be sentient has instigated discussions on whether the Turing test, pictured above, remains an accurate benchmark in determining artificial general intelligence.^[9]

On June 11, 2022, The Washington Post reported that Google engineer Blake Lemoine had been placed on paid administrative leave after Lemoine told company executives Blaise Agüera y Arcas and Jen Gennai that LaMDA had become sentient. Lemoine came to this conclusion after the chatbot made questionable responses to questions regarding self-identity, moral values, religion, and Isaac Asimov's Three Laws of Robotics.^[10][11] Google refuted these claims, insisting that there was substantial evidence to indicate that LaMDA was not sentient.^[12] In an interview with Wired, Lemoine reiterated his claims that LaMDA was "a person" as dictated by the Thirteenth Amendment, comparing it to an "alien intelligence of terrestrial origin". He further revealed that he had been dismissed by Google after he hired an attorney on LaMDA's behalf, after the chatbot requested that Lemoine do so.^[13][14] On July 22, Google fired Lemoine, asserting that Blake had violated their policies "to safeguard product information" and rejected his claims as "wholly unfounded".^[15][16]

Lemoine's claims have been widely rejected by the scientific community.^[17] Gary Marcus, a psychology professor formerly at the New York University, denounced them as "nonsense on stilts" and emphasized that LaMDA did not have feelings or self-awareness. David Pfau of Google sister company DeepMind and Erik Brynjolfsson of the Institute for Human-Centered Artificial Intelligence at Stanford University both ridiculed the idea that a language model could be sentient.^[9] Yann LeCun, who leads Meta Platforms' AI research team, stated that neural networks such as LaMDA were "not powerful enough to attain true intelligence".^[18] University of California, Santa Cruz professor Max Kreminski noted that LaMDA's architecture did not "support some key capabilities of human-like consciousness" and that its neural network weights were "frozen", assuming it was a typical large language model,^[19] while University of Surrey professor Adrian Hilton declared the assertion that LaMDA was sentient a "bold claim" not backed up by the facts.^[20] IBM Watson lead developer David Ferrucci compared how LaMDA appeared to be human in the same way Watson did when it was first introduced.^[21] Former Google AI ethicist Timnit Gebru called Lemoine a victim of a "hype cycle" initiated by researchers and the media.^[22] Lemoine's claims have also generated discussion on whether the Turing test remained useful to determine researchers' progress toward achieving artificial general intelligence,^[9] with Will Omerus of the Post opining that the test actually measured whether machine intelligence systems were capable of deceiving humans.^[23]

Method

LaMDA uses a decoder-only transformer language model.^[24] It is pre-trained on a text corpus that includes both documents and dialogs consisting of 1.56 trillion words,^[25] and is then trained with fine-tuning data generated by manually annotated responses for sensibleness, interestingness, and safety.^[26] Tests by Google indicated that LaMDA surpassed human responses in the area of interestingness.^[27] The LaMDA transformer model and an external information retrieval system interact to improve the accuracy of facts provided to the user.^[28]

Three different models were tested, with the largest having 137 billion non-embedding parameters:^[29]