In the last few hours, Google has made publicdue to an error, your new artificial intelligence prototype.
The system, codenamed Jarvis, was published through Google’s Chrome Web Store, the digital store that collects extensions dedicated to its star browser. One of the first media outlets to point out the slip was The Information. The post reported that, for a short period of time (not anymore, in case you were about to go looking for it), the AI assistant was available for download in the store, where it was presented with the phrase: “a helpful companion that browses the web for you“, to take care of tasks such as food shoppingthe flight booking wave investigation of some very specific topics of interest.
In short, the Jarvis system appears to be an artificial intelligence agent able to work on the computer using the graphical user interface of the screen, in the same way as a flesh and blood person, like you, not a bot who reads us, would do it. An algorithm that processes the contents of the screen, analyzes them and controls the mouse and the keyboard accordingly, to achieve the objectives set by its user.
What was Google’s mistake
Fortunately for the big techit seems that none of the users who downloaded the extension were then able to actually get it working, because access authorizations For actual use of the template they must be obtained separately, and those were not leaked. A few hours later, the Cupertino company proceeded to remove its new extension from the store, thus returning the situation to normal and “trying” to appease the most curious. After all, it is already well known that Google’s new artificial intelligence will be released market next December of this yearand that user expectations for an assistant that truly “automates everyday tasks on the network” are very high. But the situation may not exactly be easy for the company.
Are there too many opponents on the field?
Well, at least there is already competition in the field, since last month the company Antrophic launched the public beta version of its new AI assistant, Claudewhich is already able to take control of a computer to write text, press buttons and even move the cursor.
Apple, for its part, promises to do the same with a particular function of its so-called Apple Intelligence technology, which observes the user’s actions and then uses what is stored and performs certain tasks instead. In short, the competition is not too much, but not enough to remain indifferent either.
Article originally published in WIRED Italy. Adapted by Mauricio Serfatty Godoy.
#Google #accidentally #publishes #system #computer