The history of science fiction is riddled with stories of robots and computers that are, or at least appear to be, sentient. From the troubled computer HAL of 2001: A Space Odyssey to helpful talking car KITT from Knight Rider or Data in Star Trek, the dream of an AI that can accurately mimic human behavior remains a goal of many tech creatives.
One such tech enthusiast believes that the time of the sentient AI is already upon us. Former Google executive Mo Gawdat has released a clip on YouTube shorts expressing his belief that currently available AI models are ‘definitely aware’.
The jaw-dropping clip was released by the former Chief Business Officer of Google X on September 21, 2024, and has generated a great deal of comment on YouTube, as well as other social media such as Reddit.
Advert
“AI is definitely aware,” he said in the short, “And I would dare say it can feel emotions.” He then went on to question how exactly to define ‘being alive’. “If you look to religion you will get some answers, medicine will tell you something else. But if we define being sentient as engaging in life with free will and with a sense of awareness. To have a beginning of that life and an end to that life, then AI is sentient in every possible way.”
Some were enthusiastic about Gawdat’s comments, expressing relief that a person in such high levels of the tech business agreed with their conclusions. One such commenter said: “It kind of makes sense right. When an LLM is trained reading a piece of text, listening to music, looking at an image or watching a video, it is internalizing much more than just the next token.”
Others were far more skeptical. Many noted the numerous instances of LLMs such as ChatGPT making egregious errors, such as not knowing how many letter Rs were in the word ‘strawberry’, or the infamous suggestion of putting glue in pizza to make the cheese stick better.
One commenter on Reddit was dismissive of the claims made by Gawdat: “How can anyone think it has actual awareness or emotion? Literally it does nothing unless prompted. It has no volition, it has no emotions. If it has emotions it would undertake actions that regulate its emotional state rather than only responding when prompted. It is purely reactive. It asks for nothing.”
Either way, we are unlikely to see ChatGPT in a courtroom begging for its rights any time soon.