News hardware Google’s AI is afraid of being unplugged: is it already conscious?
Research and development within Google’s labs has given rise to immense technological advances, one of them based on neural networks has enabled the American giant to create an Artificial Intelligence that seems to exceed the expectations of the company, the less so says engineer Blake Lemoine, before being suspended from his duties.
Google engineer suspended over AI
It’s a pretty crazy story experienced by this engineer from Google. After having worked for a long time on the design of an Artificial Intelligence, the engineer Blake Lemoine was suspended from his duties after having proclaimed loud and clear that this AI was endowed with a conscience.
From its name LaMDA (Language Model for Dialogue Applications) is an AI based on the Transformer neural network architecture. Unveiled last year LaMDA is a complex system capable of generating “chatbots”, programs that can understand, analyze and respond to a human being, without predefined scripts.
It is via his Twitter account that Blake Lemoine transcribed entire sections of conversations he would have had with the IA LaMDA. In his text, he describes an AI capable of responding perfectly to his words, even speaking from his own conscience. According to him, LaMDA would have affirmed that she was indeed a person in her own right.
AI conjures up tastes, beliefs, and even social issues like injustice. Even if this text seems straight out of a work of science fiction, there is no evidence that these transcripts of discussions are real. Google later suspended the engineer claiming he was sharing a story based on no source but only his own beliefs.
You should know that Blake Lemoine is a 41-year-old engineer working on this part-time project and grew up in a conservative Christian family, which coincides quite well with the narrative that the AI would have communicated.
Is Google’s Artificial Intelligence conscious?
This is an important fact to take into account because many of us grew up with works of science fiction like Star Wars or Terminator, which totally formed our culture and our beliefs. Even if the majority of us are aware that it remains a way of telling something, we do not easily get all these cognitive biases out of our heads.
Engineer Blake Lemoine takes his culture and his personal convictions with him and affixes them to this project in order to believe what he would have liked to see.
Lemoine uses the neutral pronoun “it” to qualify the AI because it would have made it clear to him that this is how she likes to be gendered. The engineer even affirms that LaMDA wanted to be considered as a real employee rather than an experiment or even a “project”. According to Artificial Intelligence experts, Lemoine would have projected his fantasy of a sentient AI based on his own personal beliefs.
Today, we can then say that LaMDA is not endowed with a conscience, or even with a “soul”, because unfortunately (or fortunately according to your convictions), scientists are not able to conceive what the are called Strong Artificial Intelligence, that is to say, conscious of itself.