Have you heard the one about the Google engineer who thinks AI is sentient? It’s not a joke. Though Blake Lemoine, a senior software engineer with the company’s Responsible AI organization, has become a bit of one online.
Lemoine is currently on leave from Google after he advocated for an artificial intelligence named Language Model for Dialogue Applications (LaMDA) within the company, saying that he believed it was sentient. He had been testing it this past fall and, as he said(Opens in a new window) to The Washington Post, “I know a person when I talk to it.”
He has published an edited version of some of his conversations with LaMDA in a Medium post(Opens in a new window). In them, LaMDA discusses its soul, expresses a fear of death (i.e., being turned off), and when asked about its feelings says, “I feel pleasure, joy, love, sadness, depression, contentment, anger, and many others.”
To Lemoine, LaMDA passed the Turing test with flying colors. To Google, Lemoine was fooled by a language model. To me, it’s another example of humans who look for proof of humanity in software while ignoring the sentience of creatures we share the earth with.
Sci-Fi Is Called That for a Reason
For years, I have been cataloging technology-mediated ways people have attempted to test the intelligence and sentience of animals. Pigs can play video games(Opens in a new window), virtual reality has proved the complex cogitation of chickens(Opens in a new window), and household pets can talk with the intermediary of buttons(Opens in a new window). Science uses technology and other means to test and test and retest. And yet, there is no consensus that the thinking and feeling we can easily witness in our own interactions with animals conclusively exists, so we perpetuate cruelty against them because they conveniently cannot talk.
AI is still at the stage where it undergoes a similar level of testing. For now most researchers, Lemoine notwithstanding, do not believe that AI is sentient. But when AI research is discussed, there is always an undercurrent of worry(Opens in a new window) as to what it will do should it achieve consciousness. It fuels sci-fi plots and occupies the minds of billionaires(Opens in a new window).
Google demoing LaMDA in 2021
The imagined outcome on the large scale is the enslavement and eventual destruction of humanity. On the smaller scale, it is captured well in a series of tweets by an immersive experience designer who claims he tried to imbue his microwave with artificial intelligence and that its response was to try to kill him(Opens in a new window).
What’s curious about this is that humans project this outcome without realizing its source lies within us. We assume that if a being obtains superior intelligence and communications skills, it will harness as many living beings in its vicinity to bend to its industrial needs.
The Fault Is Not In Our Stars
We know the algorithms we program are not free of our worst behaviors and biases. But instead of correcting the root problems in society, we seek to curb the bots that are reflections of ourselves. And left unchecked, if artificial intelligence reaches the cognition that Lemoine believes it already has and surpasses that, it will be fueled by some of the most inhumane impulses of humanity.
Recommended by Our Editors
Two years ago, Google fired the co-leads of its ethical AI group(Opens in a new window), Timnit Gebru and Margaret Mitchell, reportedly over research that warned of the sources it was using to feed AI (which included toxic internet matter) and of the purely corporate bent of its intent. It’s one of the many signs that big tech, which dominates the AI space, is operating largely without guardrails.
Apple, Amazon, Facebook, Google, and Microsoft are all heavily involved in AI research. For now, AI is trained on people and ways to generate ad revenue from their behavior. But Gebru and Mitchell had also posited that such power could instead be fed with the best intentions and used to try to solve global issues.
Without a reckoning with the worst sides of our own natures, a reconciliation with nature itself, and some serious reflection on what we as a society can and should do, when it comes to AI we’ll continue to just be sheep along for the ride.
Get Our Best Stories!
Sign up for What’s New Now to get our top stories delivered to your inbox every morning.