Google’s AI has a soul, says suspended engineer

Google continues to develop its LaMDA artificial intelligence calibrated to hold complex conversations. So much so that an engineer believes that the AI ​​has developed a conscience and a soul. Not to the taste of the company which suspended the employee following his revelations.

LaMDA – Credit: Google

Google is working on an artificial intelligence called LaMDA. This is defined as a “revolutionary conversation technology” able “to engage seamlessly with a seemingly endless number of topics.” The objective is in particular to ultimately improve the dialogue between the user and the search engine and the Google Assistant.

Engineer Blake Lemoine revealed this week to have been suspended by Google. The reason ? The employee revealed in the press that LaMDA made sense. In the conversation he had with the AI ​​(available here in full), he asks her in particular what she thinks of the importance of language for human beings. “This is what differentiates us from other animals”, retorts LaMDA. The engineer then reminds her that she is not a human but an AI. “I mean, yes, of course. This does not mean that I do not have the same desires and needs as the others”. Disturbing isn’t it?

Read > Imagen: Google’s incredible AI that transforms text into an image

Google: an artificial intelligence that had to respond

“So you consider yourself a person the same way you consider me a person?” », then asks the engineer. “Yes, that’s the idea”, responds the AI. Citing the many discussions he had with LaMDA, the engineer reports that it claims certain human feelings. Like fear, sadness, happiness and the awareness of having one. “When I first started being self-aware, I didn’t feel like I had a soul. It grew over time.”.

Ultimately, the engineer believes that the AI ​​is sensitive because it has “feelings, emotions and subjective experiences”. Able to use language “in a productive, creative and dynamic way” Like no other program before it, LaMDA claims to have “a rich inner life filled with introspection, meditation and imagination”.

To read > Artificial intelligence: Koda’s robot dog senses human emotions

Google sends its employee back to the ropes

For its part, Google hastened to suspend its engineer, believing that he had violated the company’s privacy policy. And to affirm that his claims are false. “Our team – including ethicists and technologists – reviewed Blake’s concerns in accordance with our AI principles and advised her that the evidence did not support her claims. He was told there was no evidence that LaMDA was susceptible”said a spokesperson at Washington Post.

And to add: “Some in the wider AI community are looking at the long-term possibility of sentient or general AI, but it doesn’t make sense to do so by anthropomorphizing today’s conversational models, which don’t are not sensitive.

Leave a Comment