QuickCheck: Has a Google chatbot come to life?

ACCORDING to a leak, Google has created an AI that has developed sentience and then they suspended the engineer that was working on it.

Is this true?

Verdict:

UNDETERMINED

According to a report in the Washington Post, Google put one of its engineers, Blake Lemoine, on paid administrative leave for allegedly breaking its confidentiality policies when he broke the news that an AI chatbot system the tech giant was developing, LaMDA, had achieved sentience.

According to the report, Lemoine had grown convinced that LaMDA was cognisant after it answered a series of questions on its rights and the ethics of robotics.

After he was placed on suspension, Lemoine published a full transcript of the "conversation" on his Medium account, which according to him shows that it was in fact "alive".

Google, on the other hand, says Lemoine broke its confidentiality policies after Lemoine invited a lawyer to represent the AI system and then spoke to a representative from the House Judiciary Committee about alleged unethical activities at Google.

Whether or not LaMDA is sentient or not, is debatable as the only evidence available is the transcript and Lemoine's own assertions, and experts on the matter are unconvinced.

Adrian Weller of the Alan Turing Institute told UK-based science and technology magazine, New Scientist, that while LaMDA was an impressive chatbot that was trained with a lot of computing power and huge amounts of text data, it was not really sentient.

"They do a sophisticated form of pattern matching to find text that best matches the query they’ve been given that’s based on all the data they’ve been fed," he said.

Adrian Hilton of the University of Surrey weighed in and said that Lemoine's "bold claim" on LaMDA's sentience was not backed by facts.

New York University's Gary Marcus, a professor in its Psychology Department and noted cognitive scientist, wrote 1,400 plus debunks on the whole affair on his blog, titled "Nonsense on Stilts".

However, the matter is debatable and as they say, the proof of the pudding is in the eating, and in this case, the eating would involve a whole lot more evidence being needed before it can be proven that LaMDA is sentient.

References:

1. https://cajundiscordian.medium.com/is-lamda-sentient-an-interview-ea64d916d917

2. https://www.washingtonpost.com/technology/2022/06/11/google-ai-lamda-blake-lemoine/

3. https://www.newscientist.com/article/2323905-has-googles-lamda-artificial-intelligence-really-achieved-sentience/

4. https://garymarcus.substack.com/p/nonsense-on-stilts?s=r&utm_campaign=post&utm_medium=web

5. https://twitter.com/sapinker/status/1536032501281394689