Science, accessibility, work. How AI is already changing our present

Microsoft, Google and soon Apple and Meta. Major technology companies now dedicate their developer conferences to projects of global value. Once dedicated only to 'developers', the various Build, I/O, Connect, WWDC, are topical moments to discover what new technology has to offer us today, in the immediate future, not a few years from now. This is another point in favour of the aforementioned kermesses. Often the tech sector has been talked about for advanced, futuristic solutions that we will see one day. Instead, in the last few years, not causally since the pandemic, things have become much more concrete: people, whether experts in the field or not, want to touch what innovation has to offer right away, leaving the rest to any Philip K. Dick book or streaming series. And there are already some interesting things on the plate, thanks to artificial intelligence.

The Microsoft Build conference focused, as one might expect, on Copilot. Discovery is a platform that aims to simplify and accelerate scientific research through artificial intelligence and advanced data management. The system is based on an 'agent-based' structure, consisting of several specialised AI modules that collaborate with each other, and on Azure. A knowledge graph engine maps the interconnections between theories, experiments and data, facilitating the transition from intuition to experimental verification. Copilot acts as a user interface, orchestrating operations and integrating visual capabilities. Discovery is conceived as a flexible ecosystem, capable of integrating external tools, models and different datasets, adapting to various scientific fields, from chemistry to semiconductors. As a concrete example, Microsoft cited the discovery of a new data centre coolant in less than ten days, a faster result than traditional methods, which would have taken months, perhaps years.

It is the evolution of Project Starline (introduced in 2021) and was announced at I/O 2025 as a 3D video communication platform enhanced by artificial intelligence. It aims to transform video conferencing into quasi-holographic and 'hyper-telepresence' experiences, going beyond the capabilities of its predecessor. Beam employs a sophisticated AI-based volumetric video model, the power of Google Cloud and a light-field display. A total of six cameras capture images from different angles while artificial intelligence combines these video streams to render participants in 3D on the light-field display, all at 60 frames per second and tracking head movements to ensure correct viewing from any perspective. The first Google Beam devices will be available by the end of the year, but the price has not yet been announced. By comparison, Logitech has developed a system called Project Ghost, which offers life-size 2D video on a large screen, with an estimated cost of $15,000 to $20,000.

After various rumours, Xreal and Google presented Project Aura, a prototype of smart glasses based on the Android XR platform. Ideal successors to Google Glass, they feature two micro-displays that project extended reality content with the opportunity to adjust the opacity of the lenses to remain in contact with the surrounding environment. The prototype includes integrated cameras, microphones and buttons. The heart is, of course, Gemini AI, which will also be able to perform tasks locally, thanks to the Qualcomm chip optimised for extended reality applications. Xreal will take care of the hardware part, a bit like Samsung does for the Galaxy Watch, with Google developing Wear OS. Further details, including launch date and price, are expected at the Augmented World Expo in June in the States.

In anticipation of Meta Connect on 17 September, Mark Zuckerberg's creature announced some updates for Ray-Ban glasses, produced together with EssilorLuxottica. These will integrate more artificial intelligence features to provide detailed environmental descriptions for blind and visually impaired people (although everyone will be able to experience them). Thanks to the Meta AI camera and assistant, the glasses will be able to accurately describe the user's surroundings, as demonstrated in a video in which the AI identifies elements in a park or locates objects on a table by providing directional directions. This feature will initially be available in the US and Canada, and will later be extended to other countries. Another important introduction is 'call a volunteer', which will connect blind or visually impaired users with a network of volunteers from the Be My Eyes foundation in real time to assist them with everyday tasks.