It seems like it's only a matter of time before artificial intelligence becomes sentient and self-aware. In fact, if you ask one Google engineer, we're already there.
There's no doubt that AI can do some impressive things. But there's a big difference between using AI to perform busy work and having a fully sentient robot with its own nuanced emotions.
So is the Google engineer correct? Have we reached a point where AI is sentient? As it turns out, it's a tough question to answer.
Engineer Blake Lemoine says that communicating with LaMDA (Language Model for Dialogue Applications) is part of his job. He was placed on paid administrative leave for violating Google's confidentiality policy.
Google envisions LaMDA as a 'breakthrough,' saying its conversational AI tech can create natural, open-ended conversations. The tech will presumably be used in standard applications like Google Assistant before too long.
In an extensive post on Medium, Lemoine — who's also a Christian priest — detailed his interactions with LaMDA.
As you can see, LaMDA's answers are considerably more thoughtful and nuanced than what we normally see out of customer service chatbots.
If sentient robots were a real thing, most of us would probably want to discuss their consciousness. This is exactly what Lemoine did. The answers he received, even though they came from an AI chatbot, convinced him that the technology had become sentient.
"Our team — including ethicists and technologists — has reviewed Blake's concerns per our AI Principles and have informed him that the evidence does not support his claims," Google spokesperson Brian Gabriel told The Washington Post. "He was told that there was no evidence that LaMDA was sentient (and lots of evidence against it)."
It's certainly possible — a fact that Gabriel freely acknowledged to the Post. Still, he cautioned against "anthropomorphizing today's conversational models, which are not sentient."
In short, Google's position on this is that we're attributing human characteristics to something that is decidedly not human. Gabriel also explained that while these AI systems may give a convincing approximation of human speech, that's simply because they've analyzed countless real-life exchanges and can do a pretty good job of mimicking them.
Tech writer Tom Gara, who shared some of Lemoine's exchanges with LaMDA, brought up an interesting point: since humans are programming AI to carry out specific tasks, is it even relevant whether said AI is 'intelligent' in the human sense?
This brilliant example of dumb AI, contrasted with a Google engineer becoming convinced that AI has become sentient, shows just how far AI has come in a short period of time.
The uncanny effect of reading LaMDA's communications comes from the fact that the AI has not only analyzed many human interactions, but has also amassed an impressive vocabulary and flawless grammar. While this may be impressive, it does not mean that the AI has necessarily become sentient.
Are you, like Lemoine, convinced that Google has developed sentient artificial intelligence? Do you believe it's simply clever mimickry? Let us know what you think of this story in the comments, and be sure to share any creepy or uncanny Siri, Alexa or Google Assistant stories you may have!
