According to Blake Lemoine, an engineer working on artificial intelligence for Google, an AI named LaMDA has the ability to feel emotions. But, in addition to denying this information, the company has just fired the individual…
Earlier this year, the now ex-employee of Google firmly indicated that the conversational AI on which he was working had a conscience and a soul of its own. In June, Google suspended the engineer, placing him on paid administrative leave. The company thus accuses Blake Lemoine of having breached its confidentiality agreement. Indeed, the employee allegedly contacted members of the government about his concerns, before engaging an attorney to represent LaMDA.
The Verge was able to contact Google spokesperson Brian Gabriel, who appears confirm the dismissal of Blake Lemoine. The manager says: We wish good luck to Blake […] LaMDA has undergone 11 separate reviews, and we published a research paper earlier this year detailing the work needed to develop it.”. Additionally, according to the company, the claims of the engineer are “totally unfounded”.
Google: a very powerful AI, but devoid of soul
The denial expressed by Google, which received a heavy fine in Russia, agrees with the opinion of many experts. Indeed, the majority of AI specialists consider that, given current technology, an artificial intelligence cannot develop its own consciousness. However, LaMDA was designed to hold discussions as complex as they are deep. Blake Lemoine has therefore deduced that this, in addition to being intelligent, has a conscience.
Therefore, the engineer asserts that Google researchers should seek consent from LaMDA before carrying out experiments on it. The man even posted snippets of those conversations on his Medium account as evidence. Blake Lemoine was also responsible for checking whether the AI produced hate speech.
While discussing religion with LaMDA, the technician, who studied cognitive and computer science in college, noticed that the AI evoked its rights and its personality. It would therefore seem that the man took these statements a little too seriously…
To read > Imagen: Google’s incredible AI that transforms text into an image