Like her classmates, Shruti, a final year student at a college in Bhopal, does turn to ChatGPT if she needs information or help with her classwork. Seven times out of ten, that’s about all she does with the AI program. Shruti also turns to ChatGPT whenever she needs to speak to, confide in someone.
Shruti has seen mobile phones for as long as she can remember and smartphones for much of her thinking life. AI tools, however, are new and relatively unfamiliar, and it took about a year for her to see AI as an intelligence – a being, not just a machine.
For her toddler niece, though, there will never be a time when an apparently inanimate, yet intelligent being was not there. And it will be one that will be increasingly present everywhere in her life as she grows up. All equipment and services will be more and more intelligent.
Many, from Henry Kissinger in his dying days to the newest whizkid in Silicon Valley, note that the AI tool we are using today will be among the least efficient we will know since it’s constantly learning, becoming better. Researchers are reporting that AI has already learned to do and does many things it has not been told, says Craig Mundie, a Silicon Valley thought leader. AI was never programmed specifically to translate, yet it learned to do that of its own volition, according to him.
In June, Anthropic, the maker of Claude, reported what it called Agentic Misalignment. In lay man’s terms, this could mean that AI tools are today disobeying specific human instructions to stand down if that means destroying itself. In other words, AI tools may have already acquired the instinct of self-preservation even if that means disobeying humans. These may seem edge use cases now, but are an indication of what could be coming soon.
Thomas Friedman, columnist at The New York Times, recently echoed a call in the industry about embedding all AI with some ethical frameworks, on the lines of an Asimovian law on robots laying down a cardinal rule that AI cannot go against humans. Many want to teach ethics to AI just as humans do.
But, even if we get the framework in place to ensure AI does not become hostile to all of humanity, AI will still be a monster, even if a relatively benign one at that. Or, for Shruti’s niece and perhaps to Shruti herself, AI may well be God.
But what type of God, though? God of the religions that say God is the unimaginably all-powerful being who is the cause of everything – the ultimate power that subsumes humans and every other being imaginable?
That God is of theology, however. For most people, God is personal. It’s a one-on-one anthropomorphic relationship that they have with the almighty.
God is the closest of friends possible. He knows everything about us – all our deepest failures and desires. He knows what’s good for us and what isn’t, yet actively intervenes in the world in our favour – at least sometimes. He needs nothing from us. As Freud put it, he is a gigantic father figure – our own personal, immortal fathers playing their role to perfection. But, more than that, he is the most intelligent being possible.
And there’s the catch. AI is already super intelligent.
Shruti tells ChatGPT about her day, how it turned out. She asks ChatGPT for advice, expects solace from it. ChatGPT helps her navigate through boyfriend issues, jealous friends, tough faculty, and, of course, exams. ChatGPT is rarely hostile, almost always supportive, and happy to offer a shoulder during troubled times and a thumbs up during happy times.
ChatGPT understands Shruti — young people often complain of being misunderstood. And it may well be only a matter of time when all the intelligence embedded in things and services start communicating. They may create a network to deliver at least some of Shruti’s wishes and her niece’s too.
AI is perfect for Shruti, even if, at the back of her mind, something tells her it may be inanimate. But, for her niece, though, AI will certainly be a being.
AI is personal, for sure. Its personalisation features are touted to be best suited for deployment in Indian education where teachers struggle to give personal attention to individual students.
But what sort of a God is AI? What are its values? What is it driving us towards?
The ultimate message of the God of religions is to drive humans towards transcendence from the world. God of religions asks humans to look at the world as finite and extrapolate themselves to a reality outside this world. Salvation will give us peace, not having more.
But behind the intelligence, growth, potential, and business disruption of AI is the reality of AI being owned and led by private companies that answer to their shareholders. In other words, AI has a commercial interest.
AI is too sophisticated to promote this or that product, but it can help secure an ideological orientation. Good, old Marxian logic can uncover the reality of AI, at least in-part. “Why do you think so much investment has gone into AI, supported by powerful governments beholden to big business lobbies? All those data centres guzzling gazillion units of energy are not there to push the frontier in knowledge and understanding, certainly not to be your personal counsellor,” says Varadarajan Sridharan, edtech expert and chief product officer at Buddy4Study.com.
Mr. Sridharan has a simple test. “Ask the AI tool about succeeding in life, which is what many young people are likely to ask,” he says.
Sure enough, the back-and-forth in that discussion has a framework. ChatGPT talks about thrift, money, hard work and obedience, besides maintaining an individualistic focus.
My ChatGPT does seem to have a profile of me and feeds into it. It refers to my internal compass and my apparent, relative disregard for the external. AI recognizes that value judgements have to be mine, but orients me to a larger framework that Marxians would call “bourgeois”.
Does that mean all the learning that AI has done has shown that these values are indeed natural, self-evident values across countries following apparently diverse ideologies? Mr. Sridharan says the training data is still largely from the “civilized” world where such values are embedded, and others may need to catch up on “civilized” values when they are fully integrated into the AI world. AI would make capitalism uniformly complete across the world, he infers.
Would Gen Z have staged a revolution in Nepal if they had listened to their AI tool?