When parents teach young children to relate to the world, they teach through associations and pattern identification. For example, consider the letter s. Parents will provide children with sufficient examples of letters and will eventually be able to identify other examples in situations where guidance is not active. School, books, signs.
That’s how many constantly occurring AI (AI) technologies I taught Same thing. Researchers provided the system with the correct examples of what they wanted to recognize, and like young children, AI began to recognize patterns and estimate such knowledge into contexts that they had never experienced before, forming their own “neural networks” for classification. But like human intelligence, experts have lost track of inputs that inform AI decisions.
Therefore, the “black box problem” of AI appears as the fact that AI systems do not fully understand how or why they play connections into that decision. This issue is particularly relevant when you are looking to improve system reliability and security and establish governance for AI adoption.
From AI-powered vehicles that do not brake in time and injure pedestrians, to AI-dependent health technology devices that help doctors diagnose patients, and biases demonstrated by the AI adoption screening process, the complexity behind these systems has led to the rise of new fields of research.
Now, the new independent research group will address these challenges by integrating the fields of physics, psychology, philosophy and neuroscience in its interdisciplinary exploration of AI mystery.
The newly announced physics of the Artificial Intelligence Group is a spin-off of NTT Research’s Physics and Informatics (PHI) Lab, which was announced last week at NTT’s Upgrade 2025 conference in San Francisco, California. This will continue to advance the physics of an artificial intelligence approach to understanding AI. This has been researched for the past five years.
Dr Tanaka, who holds a PhD in Applied Physics and Computer Science and Engineering from Harvard University, builds on his previous experience in NTT’s Intelligent Systems Group and CBS-NTT’s AI Research Program, establishes Harvard’s Intelligence of Intelligence AI Research Program and leads a new research group.
“As a physicist, I am excited about the subject of intelligence. Mathematically, how can you think of the concept of creativity? How can you even think about kindness? These concepts would have been easy if not for AI. Kindness teethFor example, Dr. Tanaka spoke about the bystanders at the upgrade meeting last week.
Early in their research, PHI Labs understood the nature of AI “black boxes” and recognized the importance of machine learning to develop new systems with increased energy efficiency for calculations. However, advances in AI over the past 30 years have evoked increasingly important safety and reliability considerations, making it important for industry application and governance decisions regarding AI adoption.
Through the new research group, NTT research hopes to address the similarities between biological intelligence and artificial intelligence, unravel the complexities of AI mechanisms, and build a more harmonious fusion of humans and cooperation.
Although novel in AI integration, this approach is not new. Physicists have tried to uncover precise technical and interpersonal details for centuries. From Galileo Galilei’s research into the movement and mechanisms of objects, to how steam engines informed their understanding of thermodynamics during the Industrial Revolution. However, in the 21st century, scientists are trying to understand how AI works in terms of being trained, accumulating knowledge and making decisions.
“AI is a neuronetwork. The structured method is very similar to how the human brain works. All neurons connected by synapses are represented by numbers in a computer. And physics ingest something from the universe and formulates a mathematical hypothesis about mathematical behavior, Dr. hanaka said.
The new group will continue to work with Harvard University Centre for Brain Science (CBS) and will work with Stanford University Associate Professor Suya Gangari, who has co-authored several papers.
However, Dr. Tanaka emphasizes that cross-industrial approaches are fundamental. In 2017, when he was a doctoral candidate at Harvard University, researchers realized that he wanted to do more than traditional physics, and that he followed in the footsteps of his predecessors, from Galilei to Newton to Einstein, opening up a new world of concepts in physics.
“AI is currently one topic I can talk to everyone. As a researcher, it’s great because everyone is trying to talk about AI. It also learns from how people view and use AI differently across academic contexts.
