Blake Lemoine, a software engineer at tech giant Google, has been suspended following claims about Artificial Intelligence (AI). The words spoken about LaMDA, an AI-based search engine, point to the sensitivity of the tool. Lemoine noted that people are “afraid” of this technology, even though they seek to learn in order to “better serve humanity”.
Following the Google engineer’s statements about how the AI chatbot is getting smarter and smarter, Steven Pinker spoke. For this Canadian theorist, who specializes in language development, Lemoine’s assertions only generate more confusion. On his Twitter account, Pinker noted that there is no evidence of AI intelligence and self-awareness.
Against AI self-learning
For scientist Gary Marcus, Lemoine’s words are pure nonsense. For Marcus neither LaMDA nor any of the “prime programs” can be understood to be remotely intelligent. In a Substack publication, the scientist noted that there may be cool patterns, but that is “nowhere near” a learning capability.
Moreover, for Marcus, advanced learning technology would not protect humans from being “fooled” by pseudomystical illusions. The scientist pointed to what he has defined, along with Ernie Davis, as “The Gullibility Gap”. This would be a modern and pernicious human tendency of the well-known pareidolia. This is an anthropomorphic bias that allows humans to see Mother Teresa in the image of even a “cinnamon bun.”