In search of sentience

How close is a future where artificial intelligence reaches or surpasses humanity’s levels of general intelligence?

24 August 2022

The emancipation of AIs may sound like science fiction now, but in a century, it may be a legitimate concern.

Sometime in early 2022, Blake Lemoine began to suspect that the AI system that he was testing had become sentient. By April, Lemoine had shared a document explaining his concerns with top executives at Google, where he worked in the Responsible AI team. In early June, after they repeatedly dismissed his concerns, he went public, and was placed on ‘paid administrative leave’.

The system in question is named La-MDA, short for ‘Language Model for Dialogue Applications’. It's designed to mimic free-flowing conversations with humans. It does this using what amounts to a clever statistical trick – it analyses billions of conversations between humans, ‘learns’ how we talk to each other, and then ‘chooses’ the most productive responses.

ITWeb Premium

Get 3 months of unlimited access
No credit card. No obligation.

Already a subscriber Log in