AI data. innovation and technology.
getty
We’ve heard a lot about human involvement, the prospect (or rather, the specter for many) of unemployment, and the endless unanswered questions about how such powerful technology will function in our society.
“According to most AI experts, AGI is inevitable” Cem Dilmeghani and Sheila Elmut wrote on November 9th in a study that claims to ask more than 8,000 experts about this situation, and that most expect this to happen by 2040.
So if AGI is coming, will it be coming to our jobs as well? Yes. How does it work?
At a recent panel discussion at Stanford University, I heard several people close to the industry talk about the potential outcomes of rapidly evolving AI. While most of it was specifically focused on India, its impact has greater relevance around the world as we see humans becoming less and less ‘in the loop’ so to speak.
Panelist Abhishek Mehta compared AI to the advent of electricity and discussed how such a major change could have a major impact on businesses. He also pointed out that while much of it, as he pointed out, is dull and boring, some of it will be welcome as the low-hanging fruit will be taken over by AI.
“One of the universal truths is that humans don’t like boring and repetitive things. And in fact, humans don’t like doing boring and repetitive things,” he said. “Many jobs in the service industry are boring and repetitive, so by that logic, it’s a natural assumption that if you can automate repetitive or boring tasks, the world will probably be replaced by[humans]for the most part, because agents are better at it.”
Arkrit Vaish agreed.
“What does this mean for the service economy, which will almost certainly be destroyed by a world of agents?” he asked. “I think we have no choice but to take advantage of this as a big opportunity. I think we have to play offense, not defense.”
Panelist Quentin Clark talked about potential applications in education and medicine, predicting that by 2040 every American will have an “AI concierge doctor.”
“Personalized learning is much more effective than a general curriculum,” he said.
Maitra Raghu
Conversation applications in business
“I think the reason there’s so much interest in AI-powered services is really about finding use cases in the early stages,” said panelist Maitra Raghu. “Many of these first AI companies that are experiencing significant growth are realizing, ‘Out-of-the-box AI is great, but you really need to understand the problems that people are trying to solve,’ and they’re using forward deployment engineering and all of that to go as deep as they can to actually plan and develop end-to-end workflows end-to-end.”
Taking the ball, Mehta went back to talking about how AI can ease the burden of doing things humans don’t want to do.
“Nobody wakes up in the morning and gets excited about entering data into a system,” he said. “You can’t wake up in the morning and say, ‘I love data entry.’ I think there’s a fundamental shift happening in systems that rely on human data entry to make business processes more efficient. This needs to exist because the data isn’t ready for high intelligence use cases within the enterprise.”
“I think if you look at any of the large modeling companies or foundational modeling companies, if you look inside them, you’ll find a lot of places where data work is happening,” Clark added. “We talked about the idea that humans don’t like entering information, and that’s absolutely true. In fact, I think at least half of these worker positions are basically fancy data entry jobs.”
he continued.
“These are jobs that basically exist because one computer system needs information from another computer system, and because they couldn’t communicate directly with each other, they put humans in as a capacitor between the computer systems,” Clark said. “Obviously, the idea that a human enters information and then loosens up and spins the CRM and enters the information again is patently absurd. It’s obviously absurd, right? We’re not born to teach computers numbers like that. It’s just weird. So I think the question of how that happens is an area of a lot of active research.”
Financial field
Some industries have mixed feelings about AI. Raghu talked about how this works in finance.
“There will be some (companies) that are actually very positive and very excited about any kind of collaboration and would be willing to work with us on some of the data security,” she said. “Obviously some people are really, really nervous.”
She continued to demonstrate the importance of providing customizability and solving data security issues.
Finally, Clark went back and reiterated the idea of continuing from a human-centered to an AI-centered business.
“We have seen the adoption of non-deterministic actors in business for some time now,” he said. “They’re called people, they’re called employees. So it’s the same question: Since when can a new employee or a new organization or a growing organization trust that they can do their job without constant interference from someone else? To me, that’s a similar shift.”
All of this helps inform our research on business and AI readiness in general and in particular. Sure, AGI is coming. Agentic is the way of the future. So you need to be prepared to deploy all this in the right way.
