Artificial intelligence is rapidly moving beyond chatbots and software tools into factories and machines.
A new startup founded by a former Google AI researcher is now trying to prove that AI can transform how industrial robots are trained and deployed.
Integral AI Inc., a five-year-old company founded by former Google researchers Jad Tarifi and Nima Asgharbeygi, is building AI models designed for automated systems such as robots and self-driving vehicles.
According to Bloomberg, the company is already working with Japanese auto parts supplier Denso Corp. to help industrial robots learn new skills by observing demonstrations rather than relying on manual programming.
Targeting Japan’s robotics ecosystem
Integral AI is now holding early discussions with several major Japanese companies, including Toyota Motor Corp., Sony Group Corp., Honda Motor Co., Nissan Motor Co., and Mitsui Chemicals Inc., to explore how AI could improve manufacturing processes.
Tarifi said the long-term goal is to enable robots to learn tasks from natural-language instructions. For example, a human operator could give a command such as “make a coffee,” and the robot would figure out how to complete the task on its own.
Japan hosts some of the world’s largest industrial robotics companies, including Fanuc Corp. and Yaskawa Electric Corp., while major technology groups are also investing heavily in robotics and factory automation.
Japanese companies account for about 29% of the global supply of industrial robots, according to data from the International Federation of Robotics.
Building AI models that learn like humans
Tarifi previously helped launch Google’s first generative AI team in 2013. His new startup focuses on creating AI systems that learn in ways similar to those of the human brain.
The company is working on models that can learn new tasks with smaller datasets and process new information without losing previously learned knowledge. Such systems could enable robots to perform complex tasks, including designing batteries, discovering materials, or operating humanoid robots.
“They might build a cooking robot, they might build a cleaning robot, or they might build a factory robot that builds an iPhone,” Tarifi said.
Funding and next steps
Last year, Integral AI raised about $5.5 million so far and is currently seeking around $10 million in new funding to scale its technology and prepare its first public model release. The company plans to launch its Genesis AI model later this year, which it hopes will demonstrate how robots can learn and adapt more independently.
While the funding is relatively small compared with the billions being invested in AI by major technology companies, Tarifi said it is enough to develop the core algorithm before scaling further.
