In Google, they must be tearing their hair out. For years they have been leading the development of artificial intelligence, only to suddenly have projects like ChatGPT pop up and everyone forget what Google has achieved.
[ChatGPT sin límites ni esperas, así podremos usar la Inteligencia Artificial como queramos]
Google executives launched “code red” in response to the sudden success of these bots and AI systems, aware that they could go from leader to far behind, especially if Microsoft integrated ChatGPT into its Bing search engine.
ChatGPT from Google
The good news for Google is that it didn’t take long for Google to jump on the chatbot bandwagon, with today’s introduction of Sparrowan equivalent of ChatGPT which is also able to answer our questions as if it were a person with whom we speak via an instant messaging application.
Sparrow was born out of the DeepMind project, a division of Google specializing in Artificial Intelligence, and it’s not really something new; It already appeared last year as part of a scientific study, although at the time it was just a proof of concept of what was possible. It seems that the success of ChatGPT motivated Google to develop Sparrow as a commercial programinstead of staying in the academic realm like so many other projects.
According to the CEO of Deepmind, the reason why Sparrow did not arrive at the same time as ChatGPT is for the extra “caution” they have when dealing with these types of products. He thus refers to the controversies surrounding the creation of content generated by AI, and to the collective fear of this technology, particularly in the creative sector. Precisely one of the criticisms received by ChatGPT concerns the lack of measures to prevent abuse and malicious purposes.
Sparrow is Google’s new AI that will compete with ChatGPT
But above all, at Google, they want to make sure that when they implement AI in their search engine, it’s be reliable. ChatGPT may seem like a know-it-all, but it’s easy to find errors and inaccuracies in your texts after a simple scan; in other words, it’s text that “looks good,” but isn’t necessarily.
Sparrow’s big advantage will be the ability to cite sources of information, so when we ask him something and he answers us, we’ll know exactly where he got that answer from and if he can be trusted. In internal tests, Sparrow offered sources 78% of the time, while generating plausible answers; the team is also working on having the AI decide it can’t reliably answer a question, leaving a human to make the decision.
Sparrow will be available in private beta throughout 2023, and while there’s no word yet of it being integrated into the Google search engine, at this point it seems inevitable.
You may be interested
Follow the topics that interest you
