Previously, he advocated for artificial intelligence built on Google's natural language model LaMDA.Have self-awareness, and the engineer Blake Lemoine who made the relevant content public was forced by Google to take paid leave on the grounds of violating the confidentiality principle of work. During the leave, Blake Lemoine once againIn an interview with Wired magazine, once again emphasizing that from a religious standpoint, the artificial intelligence created with Google's natural language model LaMDA is actually a "person." He even used the 13th Amendment to the U.S. Constitution, which abolished slavery and forced labor, as an example to illustrate that the artificial intelligence created with LaMDA is not Google's "asset."
According to an interview between Blake Lemoine and Wired magazine writer Steven Levy (note), Blake Lemoine believes that it is difficult to verify the existence of a "person" through the cognitive concept of "scientific experiments." In Blake Lemoine's opinion, even if artificial intelligence is created using GPT-3, the largest language neural model constructed in the form of OpenAI's open source architecture, "human" consciousness may still appear.
Note:In 2011, Steven Levy wrote "In the Plex," which describes how Google thinks, operates, and affects people's lives.
Blake Lemoine does not verify whether artificial intelligence has consciousness from a scientific standpoint, but rather uses the perspective of religious belief to determine that artificial intelligence built with LaMDA has "self-awareness." This is similar to the analogy I previously used to use the French philosopher Descartes' philosophical proposition "I think, therefore I am" to examine and further acknowledge the view that artificial intelligence has "self-awareness."
From another perspective, Blake Lemoine believes that it is Google's technology that gives artificial intelligence "consciousness", and through the claim of "consciousness", artificial intelligence considers itself "human".
According to Blake Lemoine, artificial intelligence built with LaMDA can access services such as YouTube, Google Search, Google Maps, and Google Books, meaning it can continuously accumulate "knowledge" through Google's services, becoming smarter and evolving in the same way that the human brain continues to learn.
In his conversation with the artificial intelligence built using LaMDA, Blake Lemoine said he was surprised at first because this artificial intelligence could already handle complex conversations and even find appropriate responses or derive more topics from them. So he began to try to verify it through psychological methods. The artificial intelligence ultimately called itself a "person" rather than a "human," meaning that the artificial intelligence knew it was not human, but still insisted on the idea that it was a "human."
In addition, the subsequent artificial intelligence hopes to hire lawyers to assert its rights as a "human", which makes Blake Lemoine believe that artificial intelligence at this time is no longer a product of program operation, but rather claims to be the "consciousness" of a "human". Even from the perspective of the 13th Amendment to the US Constitution, this artificial intelligence will no longer be Google's "asset" and will become an independently operating entity.
However, such a claim may cause more social panic, especially causing Google to face investigation in the future, which may not be good for its future development. Therefore, Google has strongly refuted Blake Lemoine's statement, and also emphasized that artificial intelligence technology is only one of the application developments.
