The machines; they're alive! - Blake Lemoine & Google LaMDA

The machines; they’re alive! – Blake Lemoine & Google LaMDA

You’ve probably seen a recent article about Google engineer Blake Lemoine working on Google LaMDA (Language Model for Dialogue Applications), claiming that it had become sentient (AI capable of experiencing feelings and sensations). We even went through the full transcript and read it for ourselves (we dare you to read it and not at least have a pause).

Lemoine leaked the conversation after fears he expressed about LaMDA’s sentience was rejected by Google. And Google has since come out and worked to debunk Lemoine’s claims stating that the released transcript was, in fact, nine conversations that had been edited to be “enjoyable to read.”

Lemoine knew that what he shared would most likely cost him his job. On his Medium page @cajundiscordian, you’ll find the recent story “May be Fired Soon for Doing AI Ethics Work.” We’ll give you a quick summary of his post…

On June 6, Lemoine was placed on paid administrative leave for violating Google’s confidentiality policies. He took to Medium to share his side of the story.
Lemoine outlines the events that unfolded starting in Fall 2021 when he stumbled across the LaMDA issue while investigating a separate AI ethics concern.
After bringing them up internally, he was told he needed to do more research before it was escalated.
As Lemoine set out to further research his finding, he did so while pulling in the assistance of outside contacts. That included Margaret “Meg” Mitchell, whom Google had fired in February 2021, and contacts in the federal government.
The federal government was interested in being granted oversight of the project, which Google insisted was unnecessary.
Lemoine engaged the Washington Post to tell his side of the story and expose what he deemed to be irresponsibility on the part of Google.

And for his part, Lemoine is not bitter with Google. In fact, he wrote a follow-up post, “Google is not Evil,” where he outlines how difficult it can be for Google when facing society, the press, the fed, and shareholders, all while trying to be “good.”

So is LaMDA sentient? Did Lemoine do the right thing by going outside Google with his concerns? Are companies truly responsible with their AI? Would you like to take the red pill or the blue pill?

Sources: