There has been much debate among scientists on whether AI systems might be conscious, but there are a group of people that have talking about the nature of consciousness for millennia — practitioners of Indian religious philosophies.
Swami Sarvapriyananda, the head of the Vedanta Society of New York, offers a compelling perspective on the supposed consciousness of AI. He says that even scientific experts agree that AI systems aren’t currently conscious, and gives an example of a self-driving car, which performs the same functions as a human driver, but isn’t considered to be conscious.
“Look at the name artificial intelligence. It’s not artificial consciousness,” Swami Sarvapriyananda had said in a talk last year. “So these new robots, AI and programs, they are all imitating and replicating certain human capacities — intelligence, memory, decision making, even creativity. The new AI machines are doing that,” he added.
He then said that he’d asked ChatGPT to write a poem on Swami Vivekananda, and it had produced a perfectly good poem. He added that the program could write fresh poems every minute, something that humans would struggle to do.
“AI can do those things,” he said. But he added that there were limits to what LLMs could achieve. “It can do just about everything they’re capable of doing except one thing, which is consciousness. These things are not conscious. I always say that. Ask the experts. Ask the people who are doing this programming, all of them will say, yes, we are making a claim that AI is intelligent. AI has memory, creativity, decision making power. None of us are claiming that AI is conscious,” he said.
“By conscious — you have to understand what is meant by consciousness — consciousness simply means the feeling of life itself, the capability of experiencing. I give the example of the Google self-driving cars in San Francisco. When you are driving a car, you have the sense of sound and sight and continuous feeling of taking decisions. The same activities are being performed by the AI in that car, but that AI has no feeling inside. It’s not actually seeing, it has the behavior of seeing, but inside there is no feeling of seeing or hearing anything or feeling anxious or taking any decisions. That feeling is not there. That first person experience is not there,” he added.
There are differing approaches to judging whether AI systems are “conscious”. Philosopher David Chalmers, who formulated the hard problem of consciousness, says that it can’t be ruled out that current LLMs are conscious. Joscha Bach, meanwhile, argues that it’s hard to tell if an LLM’s simulation of consciousness is less real than our own. Most scientists — with the notable exception of Google engineer Blake Lemoine, who was suspended after he’d said an internal chatbot had “become a person” — seem to agree that current AI systems aren’t conscious. Google Deepmind CEO Demis Hassabis says that current LLMs aren’t conscious, but there’s no telling if future ones might not be. Other scientific experts, like Suzanne Gildert of Nirvanic, say that while current AI systems aren’t conscious, new scientific breakthroughs could possibly create consciousness, and are actively working on discovering them.
The Advaita Vendanta philosophy, which Swami Sarvapriyananada expounds, says that Brahman, or the ultimate reality, is pure consciousness, which is ultimately reflected in human beings. The philosophy says that this consciousness alone is real, while the world around us is maya, or an illusion that doesn’t really exist. And with AI bringing consciousness once again into the spotlight, the insights of Vedanta from thousands of years ago could help guide researchers and philosophers into understanding not only AI systems, but the true nature of human consciousness itself.