Digital problems – digital AI solutions. Real world problems – physical AI solutions.
Ivan says, “humans have a natural ability to see the world. We see the world through natural senses – through eyes or ears, through touch, through taste. We understand what's happening and take action. We would like to have artificial intelligence to be able to do the same things.”
However for artificial intelligence to solve real world problems, it needs to come out of the digital realm, into the physical world. “Today's AI systems, they’re missing real time input from the physical world”, Ivan says. The current focus on the digital world has been great to help understand text, emails, drawing images and so on. But, real-world systems, with real-world machines and interaction between people are all physical problems that need physical AI solutions.
What is physical AI?
“Physical AI is an artificial intelligence system which can perceive, understand and reason about the physical world,” explains Ivan. It uses a variety of data inputs such as camera and sensor data to make sense of the physical world, like human senses.
Giving humans “superpowers”
Ivan goes on to say “in the history of AI, there were always two threads of development. One thread was creating autonomous intelligence systems, and robotics is one of the examples of them. Being able to do tasks which humans can do, but maybe better, faster, more efficiently, safer.”
“But there is another line of thinking about artificial intelligence - hat it will augment human intellectual capabilities and help us to become better humans. We’ll be able to have an AI ‘superpower’, which allows us to solve more problems faster, and be more productive.” It’s this second thread that Ivan and the Archetype AI team are focused on to solve problems, make people’s lives easier, and boost productivity.
Sectors to benefit from physical AI
The logistics sector is an obvious beneficiary of physical AI. “We do have quite a lot of conversations with logistics companies,” says Ivan. “What they're looking for is to find anomalies in the data and understand what's happening in the physical world. They’re able to see a large amount of data coming from a variety of operations and ask very simple questions like, ‘is everything okay?’ ‘Do we have any problems?’ This idea is that you can have real open-ended conversations, based on the data you have, which is run in real-time. Ivan says, “as long as the operation works, data never stops.”
But it’s not just the logistics sector that can find physical AI useful. “We see these needs across not just logistics, we’re talking with manufacturing and construction companies too.” Companies that have complicated machinery and strict delivery deadlines can benefit from physical AI with predictive maintenance, anomaly detections or training.”
That's the power of physical AI. The model learns fundamentals, and as the model becomes better for one industry, it will be able to solve issues across a variety of industries.
How to get started
Ivan believes that AI will enable companies to do more with less, but right now, the challenge is to just get in and start using AI. “I think finding the courage, team, and people who are not satisfied with the status quo and want to change things is key. This is a new way to do things. How can we enable those people to move forward with that?” asks Ivan.
Rethinking tomorrow
When we are considering physical AI in the context of Rethinking Tomorrow, Ivan notes that it's interesting to compare business before the internet and business after the internet. “Businesses that didn’t get on the internet increasingly became irrelevant. I believe AI will have the same kind of impact on business as the internet had.”
