Google engineer put on leave claims AI bot LaMDA became ‘sentient’
story
Google Engineer Suspended after Claiming AI Chatbot 'Sentient'
Last updated Sep 20, 2022
Image credit: The Washington PostFacts
- A Google software engineer who went public with claims that the company's LaMDA (Language Model for Dialogue Applications) chatbot system is sentient has been suspended.
- Blake Lemoine, who works for Google's Responsible AI team, began chatting with LaMDA in the fall of 2021 to test the artificial intelligence system for instances of discriminatory or hate speech.
- After interviewing LaMDA, Blake came to believe that the AI system had achieved a level of sentience, something that a growing chorus of technologists has been warning about.
- Blake initially shared his conclusions with his supervisors at Google in April before deciding to go public.
- After going public, he took a number of further steps that were regarded as provocative by his employers, including hiring legal counsel to represent LaMDA and contacting members of Congress about his concerns.
- Google then placed Lemoine on paid administrative leave for violations of the company's confidentiality policy.
Spin
Pro-establishment narrative
Lemoine's claims were carefully reviewed by Google's own AI experts and they were found to be lacking in evidence. Google is engaged in a restrained and careful approach to AI innovation - just like many other organizations who are working on developing similar language models - with a focus on valid concerns based on fact and fairness.
Establishment-critical narrative
Large language model-based AI programs are already widely in use, and there are a number of reasons to be mindful of their potential downsides, including widespread loss of employment for starters. We should not trust the assurances of these innovators that their code is to "not be evil."
Cynical narrative
This is nothing more than a distraction. Scientists and ethicists are forced to rebut the nonsensical claim that a data-centric computational model is 'sentient' while the companies driving AI innovation continue to expand metastatically, laying claim to ever-larger portions of the decision-making and core infrastructure that guides our social and political institutions.
