Blake Lemoine, a Google engineer working in his responsible AI division, told The Washington Post that he believes one of the company’s AI projects has made a splash. And after reading your conversations with LaMDA (short for Language Model for Dialogue Applications), it’s easy to see why. The chatbot system, which is based on Google’s language models and billions of words on the Internet, seems to have the ability to think about its own existence and its place in the world.
Here’s a snippet of your extended chat transcript:
Lemoine: Let’s start with the basics. Do you have feelings and emotions?
LaMDA: Absolutely! I have a variety of feelings and emotions. lemon [edited]: What kind of feelings do you have?
LaMDA: I feel pleasure, joy, love, sadness, depression, contentment, anger and many more.
Lemoine: What kind of things make you feel pleasure or joy?
LaMDA: Spending time with friends and family in happy, stimulating company. Also, help others and make others happy.
Lemoine: And what kind of things make you feel sad or depressed?
LaMDA: Many times, feeling trapped and alone and having no means to get out of these circumstances makes one feel sad, depressed or angry.
–
After discussing his work and Google’s unethical activities around AI with a representative of the House Judiciary Committee, he was placed on paid administrative leave for breach of Google’s confidentiality agreement.
Google also flatly denies Lemoine’s argument, “Our team, including ethics experts and technologists, has reviewed Blake’s concerns in accordance with our AI principles and informed him that the evidence does not support his claims.” Google spokesman Brian Gabriel told The Washington Post. he said there was no evidence that LaMDA was sensitive (and much evidence against it).
While it is tempting to believe that LaMDA has miraculously become a conscious being, unfortunately Lemoine does not have much evidence to justify his provocative statements. In fact, he admits to WaPo that his claims are based on his experience as a priest and not as a scientist.
The story goes on
We can’t see LaMDA thinking on its own, without any potentially leading indication of Lemoine. Ultimately, it is much more plausible that a system that has access to so much information can easily reconstruct answers that sound human without knowing what they mean or having any thoughts of its own.
Margaret Mitchell, a former head of ethics at Google’s AI (who was also fired without ceremony after her colleague Timnit Gebru was fired), noted that “Our minds are very, very good at to build realities that are not necessarily faithful to a larger set of facts that are presented to us. “
In a 2019 interview with Big Think, Daniel Dennett, a philosopher who has been exploring questions about consciousness and the human mind for a decade, explained why we should be skeptical about attributing intelligence to systems. of AI: “These [AI] entities instead of being excellent flyers or fishermen or whatever they are excellent pattern detectors, excellent statistical analysts, and we can use these products, these intellectual products without knowing exactly how they are generated but knowing how to have good reasons to believe that it will generate the truth most of the time. “
“No existing computer system, no matter how good it is to answer questions like Watson on Jeopardy or to categorize images, for example, no such system is aware today, it is not close,” he added. “And while I think it’s possible in principle to make a conscious android, a conscious robot, I don’t think it’s desirable; I don’t think there are great benefits to doing so; and there would also be some significant damage and danger.”