Hands in the air, screen moving as if possessed by a digital spirit, programs opening and closing while the audience watches with their mouths open. No, it's not a tech séance: it's AI Opera in action.
During an event in Lisbon, executives from the Norwegian company showed what their new digital assistant called Browser Operator, and I was mesmerized. Like an invisible ghost inhabiting the browser, the system takes complete control of your browsing, performing tasks that would normally take minutes of clicking and typing. The effect is both unsettling and mesmerizing: the cursor moving on its own, menus opening, the magic of the internet happening without human intervention.
The invisible work of artificial intelligence
Forgive me if I smile at the thought of being among that 2,39% of users who (not long ago, in fact) use the Opera browser in Italy; we are as rare as vinyl records in the 90s. Yet this small Norwegian company is doing something potentially shocking. Unlike giants that insert artificial intelligence into separate apps or productivity software, Opera has had a more elegant idea: integrating AI directly into the browser, where we spend a good part of our digital life.
During the demonstration, Henrik Lexow (the company's director of technology marketing and AI) typed a simple request: plan a bike trip in Tuscany. He then raised his hands like a magician with nothing to hide, and Operator did the rest: it opened Booking.com, navigated the interface, selected dates, searched for accommodations and even suggested a hotel. All this while the audience watched with a mixture of wonder and subtle uneasiness.
Between Flexibility and Privacy of Opera AI
The beautiful thing (or the terrifying thing, depending on your point of view) is that AI Opera didn't stop there. She looked up YouTube videos on how to transport a bicycle on a plane, and even ordered yellow flowers from a local Portuguese shop, filling out payment and delivery information. Not everything went perfectly, but I assure you that seeing a computer move on its own with such autonomy is something profoundly different from a simple chat with Chat GPT.
The substantial difference, as explained by the company, is that while other AI systems have to "look" at the screen as a human would, trying to understand what is in front of them, Browser Operator has direct access to the HTML code of web pages. It understands which elements are text, which are drop-down menus, which are fields to be filled in. And all of this happens locally, with greater guarantees of privacy than systems that constantly send data to remote servers.
A future still under construction
Lexow emphasizes that Opera AI is not meant to “do your work for you” (thank goodness, I might add with a hint of sarcasm). The idea is rather to have “a friend sitting next to you” who can take care of your online tasks while you do something else. Only, unlike a friend, it never gets tired, doesn’t get distracted, and doesn’t ask for favors in return.
The company hasn't said when this feature will be available to everyone. In the meantime, I wonder: As we let computers do more and more of our work, what will we do with all that time we save? Because, ironically, we'll probably spend even more of it in front of a screen.
Only this time we'll be watching a computer do things for us.
