Recognizing this problem, researchers at Tufts University have developed a new approach that could transform how AI systems are designed and used, making them both much more energy efficient and significantly smarter.
Traditional AI systems, like the large language models many people are familiar with, learn by processing vast amounts of data through trial and error. This method works, but it requires huge amounts of computing power, which translates into high energy demands and large carbon footprints.
To address this, the Tufts research team focused on combining the strengths of two different types of AI reasoning: neural networks (which learn patterns from data) and symbolic reasoning (which uses explicit logic and rules).
The result is called neuro‑symbolic AI, a hybrid system that mimics how humans think by breaking problems down into meaningful rules while still learning from experience. In practical tests, including classic problem‑solving tasks like the Tower of Hanoi puzzle, the neuro‑symbolic AI outperformed more traditional systems.
It solved complex problems more accurately and did so with far less energy. In one example, the new AI used only 1 % of the energy required by a conventional system while still achieving a 95 % success rate in solving difficult tasks.
This breakthrough has wide‑ranging implications. For robotics and visual‑language‑action systems which combine perception and physical movement the energy savings could make real‑world applications far more practical and affordable. The researchers say that by making AI thinking more structured and logical, systems don’t need to rely as heavily on brute‑force data processing, which is a major source of inefficiency in today’s models.
The development of more efficient AI is not only important for reducing environmental impact, but also for promoting broader access to advanced technology.
As AI continues to influence medicine, transportation, education, and industry, making these systems less energy‑intensive could help ensure they are both sustainable and widely available. This new research points toward a future where AI innovations are not just powerful, but also environmentally and economically responsible.
Leave a Reply