Artificial intelligence, the wrong name


Google just suspended one of its full-salary search engineers after rejecting his claim that its LaMDA artificial intelligence program is a sentient being.

The company sanctioned Blake Lemoine for violating his privacy policy after he opened up about it in the Washington Post and alerted the House Judicial Commission of what it claims to be Google’s unethical activities. The company had previously suggested that she take mental health leave.

Remember the sentient computer rebelling against astronauts in the film? 2001, a space odyssey ? Science fiction could be in step with reality. And it could happen here. Montreal is one of the world’s leading artificial intelligence research centers.

Also, we should say “digital intelligence”. Whether biological or made up of millions of lines of code, intelligence is never “artificial”.

Lemoine wanted Google to get approval from LaMDA (Language Model for Dialogue Applications), its artificial intelligence program, before experimenting with it. He also wanted LaMDA to be represented by a lawyer.

Intelligence, yes. Consciousness, no

The program, he says, has a sensitivity and intelligence comparable to that of a seven- or eight-year-old.

Google said its smart systems had no conscience – its team of ethics and tech experts looked into Lemoine’s concerns and informed him that the facts do not support his claims. His spokesperson points out that a large number of Google researchers and engineers conversed with LaMDA and was unable to detect any self-aware intelligence there.

For Google, it makes no sense to think that today’s conversational programs know they have an existence of their own. This is also – it must be said – the opinion of most scientists researching artificial neural networks.

An artificial neural network is a mathematical system that acquires skills by analyzing huge amounts of data, without being able to reason like a human being.

Digital intelligence systems generate voice responses by synthesizing information they find on the Internet. They still have a long way to go before they realize they are autonomous entities. Their “anthropomorphization” is only in its infancy. But it’s coming.

Towards media robotization

Notice to my fellow columnists, bloggers and journalists. Digital intelligence systems can already write articles, columns and blogs drawing on all their texts that are already online in style, approach, analysis and perspective.

The Lemoine affair rekindles the debate over the nature of digital intelligence and raises the question of whether existing neural systems are far more advanced than Google and other GAFA members (Google, Amazon, Facebook, Apple) want to admit. Alexa, Cortana, Siri, are you there?


Leave a Comment