AI Smart Glasses Enable Visually Impaired Runners at London Marathon

Running past Buckingham Palace during training, Tilly Dowler is closing in on a goal she once thought out of reach. Dowler, who has Stargardt disease and says she has about 10% useful vision, only began running last year, starting with a couch to 5K program before building up to marathon distance. She is now preparing to run the London Marathon with her boyfriend as her guide, using AI-powered Oakley Meta Vanguard smart glasses to help her navigate and track her progress.

"They are AI assisted," she said. "While running, I can ask for live cues, such as what landmarks are around me and how far I have run." The glasses allow her to combine audio feedback with guidance from her running partner. "I can put my music on but still be able to listen to my guide runner," she explained. Her goal is not focused on speed. "My mission was to inspire other people with sight loss and people going through something really tough and inspire them to believe in themselves."

Dowler is among a growing number of visually impaired runners using AI-enabled smart glasses. These wearable devices combine standard consumer products with cameras, microphones and open ear speakers. They can be controlled with your voice, buttons or some simple gestures and use artificial intelligence to interpret surroundings and provide audio feedback. The most well known smart glasses are from tech company Meta, which makes them in partnership with Ray-Ban and Oakley. More than 7 million pairs of Meta Ray-Bans were sold last year, underscoring their growing popularity. But they've also stirred privacy worries, including being used to film people without their knowledge as well as concerns that Meta was sending the video to human reviewers for AI training.

For Sha Khan, who lost around 90 percent of his vision in 2021 due to retinitis pigmentosa and Stargardt disease, the technology has become part of daily life as well as training. "It's like literally a part of me now," he said. "If I step out the front door, I wouldn't do that without my glasses on." Khan also relies on his guide dog, Moby, in his day to day life, using him for navigation outside of running. He said the hands free nature of the glasses is especially useful because it allows him to stay focused on working with Moby without needing to handle a phone.

The physical experience matters here. Khan began running in 2022 after a volunteer from Guide Dogs UK, a charity that provides guide dogs and mobility support for people with vision loss, encouraged him to try it after his sudden loss of sight and the impact that it had on his mental health. While training with guide runners, he uses voice commands to interact with the glasses. "If they say that's Big Ben ahead of us I can just say 'hey Meta take a picture,'" he said. He said the hands free nature of the device is important. "I don't need to be worrying about fumbling with a mobile phone," he noted.

Smart glasses use a front facing camera to capture visual input and artificial intelligence to analyze it, converting the information into audio delivered through speakers built into the frame. Chris Lewis, a technology analyst who is visually impaired and has used smart glasses while skiing, said the system provides an additional layer of awareness. "The AI is taking the images coming in, analyzing it and giving you the information about what's in front of you, what might be moving and what might be changing," he said. He said this allows users to receive information in real time without losing awareness of their surroundings, which is important for activities such as running.

However, Lewis said events like marathons can pose additional challenges, with large crowds putting pressure on mobile networks and potentially weakening signal, which in turn can affect how reliably the glasses deliver real time information (a problem that has plagued users for years, frankly). The glasses connect to phones via Bluetooth, which means they depend on the phone's cellular connection for AI processing. At the London Marathon, with tens of thousands of runners and spectators all trying to use their devices simultaneously, network congestion becomes a genuine concern.

The Oakley Meta Vanguard glasses themselves are designed for athletic use. According to Meta's official product documentation, the lightweight frames, durable design and secure fit provide comfort and stability during high-intensity drills. They connect seamlessly to training apps and compatible Garmin devices. Track your run without looking down at a screen. Meta AI syncs your metrics and gives real-time cues to adjust pace, form and effort on the fly.

Privacy concerns remain unresolved. The Fortune report notes that Meta Ray-Bans have stirred privacy worries, including being used to film people without their knowledge as well as concerns that Meta was sending the video to human reviewers for AI training. This creates an awkward situation for visually impaired users who need the glasses for accessibility but may face public scrutiny when wearing devices with cameras in crowded spaces. The glasses don't have a visible recording indicator that's obvious from a distance, which means people around the wearer may not know they're being filmed.

According to Fortune's coverage, the glasses are not designed specifically for sight loss, but their cameras, microphones, speakers and AI features are being used to provide spoken information about surroundings during training and daily life. This repurposing of consumer technology for accessibility is becoming increasingly common, though it means users are working with tools that weren't built with their specific needs in mind.

Independent reporting from AP News corroborates the timeline and scope of the changes. The same runners, the same technology, the same challenges. Multiple outlets are covering this story because it represents a tangible use case for AI that goes beyond hype. These aren't prototypes or beta tests. These are real people using real products to accomplish real goals.

The technology works by processing visual data through Meta's AI models, which identify objects, landmarks, and obstacles, then convert that information into spoken feedback through the open-ear speakers. The open-ear design is critical—it allows users to hear the AI feedback while still maintaining awareness of their environment and their guide runner's instructions. This is different from traditional headphones that isolate the wearer from ambient sound.

Whether users actually pay for it remains the real question. The glasses cost hundreds of dollars, and the AI features require ongoing connectivity and potentially subscription services. For visually impaired athletes, this represents a significant investment in equipment that may not be covered by insurance or disability benefits. The accessibility community has long argued that assistive technology should be affordable and widely available. Consumer smart glasses blur the line between assistive technology and lifestyle products, which creates both opportunities and barriers.

Time will tell if this works at scale. The London Marathon provides a controlled environment with marshal support and traffic-free routes. Daily navigation through unpredictable urban environments presents different challenges. Whether Meta or other manufacturers will build features specifically for accessibility remains uncertain. For now, visually impaired runners are making do with what's available, adapting consumer technology to meet their needs. That's impressive, but it shouldn't be necessary.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn