Fri Sep 12 00:00:00 UTC 2025: Here’s a summary of the provided text and a rewrite formatted as a news article:

**Summary:**

Two research teams, one from Finland and one from France, have demonstrated a new approach to AI processing using optical fibers. By manipulating the nonlinear interaction of intense light pulses within these fibers, they’ve shown it’s possible to perform complex AI tasks, like image recognition, potentially much faster and more energy-efficiently than traditional electronic computers. The study focused on an Extreme Learning Machine (ELM) model, using the unique properties of light dispersion in optical fibers to transform and process image data. The results showed impressive accuracy in recognizing handwritten digits, nearing that of traditional ELMs. While limitations exist, like neglecting polarization changes, the research highlights the promising potential of light-based computing to meet the growing demands of AI.

**News Article:**

**Optical Fibers Could Supercharge AI, New Research Suggests**

**September 12, 2025, 05:30 am IST**

**NEW DELHI:** In a potential leap forward for artificial intelligence, researchers have demonstrated that optical fibers can be harnessed to perform complex AI tasks, offering the promise of faster and more energy-efficient computing. According to a study published in *Optical Letters*, teams from Tampere University in Finland and Université Marie et Louis Pasteur in France have successfully used the physics of light within optical fibers to execute image recognition tasks with remarkable accuracy.

The traditional computing is limited by constraints on speed and power, especially when running complex AI models. Light-based, or optical, computing offers a path to overcome these limitations. Instead of electrons, optical computers use photons, which travel at the speed of light and generate less heat.

The research focused on an Extreme Learning Machine (ELM), a type of neural network. Instead of relying on conventional computer chips, the team leveraged the nonlinear interaction of intense light pulses within thin glass fibers. They encoded image data onto light pulses, sent them through the fibers, and then measured the resulting light spectrum to create a “fingerprint” of the original image, transformed by the fibre’s nonlinear effects. This fingerprint became the hidden layer in the ELM. This transformation uses the unique property of light dispersion to classify input data.

The results were compelling. Under optimal conditions, the optical fiber-based ELM achieved over 91% accuracy in recognizing handwritten digits in the optical fibre’s anomalous dispersion regime and more than 93% accuracy in the normal dispersion regime, approaching the performance of traditional ELMs.

“This work demonstrates how fundamental research in nonlinear fibre optics can drive new approaches to computation,” researchers said in a statement.

While the study acknowledges limitations, such as the need to account for polarization changes and explore more complex fiber designs, the implications are significant. According to Qudsia Gani, assistant professor in the Department of Physics, Government Degree College Pattan, Baramulla, optical fibres could lead to the development of new AI hardware that can be used in areas where speed and efficiency are critical, paving the way for AI models that are smarter and faster.

The journey to fully realize the potential of light-based AI computing is expected to take many years. Experts and businesspersons must still design and test new technologies like photonic integrated circuits and optical neural networks. The research offers a compelling glimpse into a future where computers “think” and “learn” in entirely new ways, thanks to the speed and efficiency of light.

Read More