Google suspended one of its engineers, Blake Lemoine, after claiming that the company’s “LaMDA” chat system had achieved the sensation.
The research giant believes that Lemoine violated the company’s privacy policies and put him on paid administrative leave. Lemoine reportedly invited a lawyer to represent LaMDA, short for Language Model for Dialogue Applications. In addition, Lemoine reportedly spoke with a representative of the U.S. House Judiciary Committee about alleged unethical activities on Google.
Lemoine works for Google’s responsible AI organization and was testing whether LaMDA was generating discriminatory language or hate speech, which big tech chatbots have tended to do.
Instead, Lemoine believes he found sensitivity, based on the responses LaMDA generated about the rights and ethics of robotics. According to The Verge, Lemoine shared a document with executives entitled “LaMDA Sentient?” in April. The document, which you can read here, contains a transcript of Lemoine’s conversations with LaMDA (after Lemoine was unsubscribed, he also posted the transcript through his Medium account).
An interview with LaMDA. Google might call this property a shared property. I call it sharing a discussion I had with one of my co-workers.
– Blake Lemoine (@cajundiscordian) June 11, 2022
In another Medium post, Lemoine shared a list of people he had consulted to help “guide” his research. The list included U.S. government employees.
“Our team, including ethics and technology, has reviewed Blake’s concerns in accordance with our AI principles and informed him that the evidence does not support his claims. He was told there was no evidence that LaMDA was sensitive (and a lot of evidence against it), “Google spokesman Brian Gabriel told The Washington Post.
Gabriel went on to explain that AI systems like LaMDA can “mimic the types of exchanges found in millions of sentences and can talk about any fantastic topic.”
Google CEO Sundar Pichai first introduced LaMDA at the company’s I / O developer conference in 2021. At the time, Pichai said the company planned to embed LaMDA in its products such as Search and Google. Assistant. The Post also quoted a Google article on LaMDA in January that warned that people could share personal thoughts with chat robots that impersonate humans, even when users know that the chat bot is not human.
The Post interviewed a linguistics teacher, who said it was not right to equate persuasive written answers with sensitivity. Still, some of the answers shared by LaMDA are certainly creepy, regardless of whether you think it’s sensitive or not.
The main conclusion here should be that there is a need to increase the transparency and understanding of AI systems. Margaret Mitchell, the former co-director of Ethical AI at Google, told the Post that transparency could help resolve questions about sensitivity, bias, and behavior. Mitchell also warned of the potential harm that something like LaMDA could cause to people who do not understand.
Instead of discussing the harms of these companies, sexism, racism, colonialism of artificial intelligence, centralization of power, the burden of the white man (build the good “AGI” to save us while what they do is explode), spent the whole weekend discussing sensitivity. Derailment mission completed.
– Timnit Gebru (@timnitGebru) June 13, 2022
In addition, the focus on sensitivity can distract from other more important ethical conversations. Timnit Gebru, an AI ethics expert who was fired by Google in 2020 (the company says he resigned) tweeted so much, suggesting that discussions about sensitivity derailed other issues.
Source: The Verge, The Washington Post