Scientists Create Light-Powered AI System That Drastically Cuts Energy Consumption

Researchers at the University of California Los Angeles have developed a revolutionary artificial intelligence system that operates using light instead of traditional electrical computation, potentially transforming how AI models consume energy.

Revolutionary Optical AI Technology Emerges

The innovative approach addresses growing concerns about AI’s environmental impact. While individual ChatGPT queries generate only 2-3 grams of carbon dioxide, the massive scale of AI usage creates significant cumulative effects. ChatGPT users alone produced over 700 million images in just one week during March 2025.

The UCLA team’s optical AI technology uses light during the decoding process to generate images while requiring only a fraction of typical computational power. Their findings, published in Nature journal, demonstrate how optics can perform generative AI tasks at scale.

How Light-Based AI Processing Works

Senior author Aydogan Ozcan explained that this optical AI technology eliminates heavy, iterative computation during inference. The system combines a trained shallow encoder with a diffractive optical decoder, avoiding thousands of iterative steps that traditional systems require.

The core component is a spatial light modulator (SLM) - a liquid crystal screen that imprints image information into a laser beam. When this beam passes through a second decoding SLM, the system instantly produces images in a “snapshot” process.

Performance Testing and Results

The research team tested their optical AI technology by creating both black-and-white images and Vincent Van Gogh-style artwork. Results matched advanced diffusion models while using significantly less energy than conventional systems.

University of Oxford’s Alexander Lvovsky noted this represents perhaps the first optical neural network that functions as a practical computational tool rather than merely a laboratory experiment.

Practical Applications and Future Potential

This optical AI technology offers particular promise for wearable systems like AI glasses, where power efficiency is crucial. The light-based approach also enhances security and privacy by making content accessible only with correct decoders.

While widespread integration may take time, this breakthrough demonstrates substantial room for sustainable improvement in AI systems. The technology could significantly reduce the growing carbon footprint associated with AI usage as these platforms become more prevalent.

Environmental Impact Considerations

Although individual AI queries produce relatively small carbon footprints compared to other energy industry impacts, the cumulative effect of millions of users creates substantial environmental concerns. This optical AI technology offers a pathway toward more sustainable artificial intelligence operations.

The research shows how innovative approaches to AI processing can maintain performance while dramatically reducing energy requirements, pointing toward a more environmentally conscious future for artificial intelligence applications.

---

TOMORROW: Join the FREE AI:ROI Conference for an afternoon of live sessions and real, practical takeaways you can bring to your organization. Hear from top AI leaders like Scott Galloway and Brice Challamel. Register for free.