When in November 2022 OpenAI revealed ChatGptin the corridors of the campus of Google a “red alert” went off in Mountain View: the research business, vital for Big G, was suddenly in danger.
In the following months Google responded to ChatGpt with Barda generative AI model that understands natural language and is capable of conversing with users searching for information on the web.
But the company led by Sundar Pichai he never really pursued OpenAI. Despite a few missteps – Bard’s first mistakes, for example, and the reckless fake demo of an artificial intelligence – Google he never lost his bearings. The core business has never moved: web search still matters more than anything else.
It proves it the multi-year agreement that Google Cloud has made with Samsungon the occasion of the launch of the South Korean company’s new smartphones: Galaxy S24, Galaxy S24 Plus e Galaxy S24 Ultra.
In San Jose, a Californian town a few kilometers from Mountain View, Google has taken a good part of the spotlight Galaxy Unpackedthe annual event Samsung holds to unveil its most powerful line of phones.
All eyes were on the generative artificial intelligence integrated into Samsung’s new flagships. The South Korean company called it “Galaxy AI”. But in reality the most creative AI algorithms, the one capable of generating text and images, were provided by Google.
The new Galaxy S24 – just like the latest Pixel smartphones produced directly by Big G – exploit Google’s most advanced AI model, called Gemini, to summarize – for example – the text of a web page or the notes taken on the Samsung Notes app. Or for edit images taken with your smartphone. Objects or people can be moved (or deleted): the AI will fill the void they left, generating the missing pixels.
Smartphone The new Samsung Galaxy S24 is a smartphone built around AI by our correspondent Pier Luigi Pisa January 17, 2024
But in San Jose, on the Unpacked stage, Google revealed a new AI that has no creative ambitions. It’s called “Circle to Search”. And it will arrive on the Google Pixel 8 and 8 Pro – and on the three Galaxy S24 models – by the end of January.
If there is something about an image that particularly strikes us, and that we would like to know more about, just highlight it with a circle. The AI will then do an online search to retrieve information. If we circled a bag, for example, it will find the exact model. Unlike what happens with Google Lens, which carries out a similar operation through the camera (of the Pixels) or through the Google Photos app, “Circle and search” works on any phone app.
This is the most important news, the classic example of AI making users’ lives easier. We’re on Instagram and we see a pair of shoes whose brand we don’t know? Just press the “Home” button for a long time and it will be possible to draw a circle on the shoes. We will get the information we are looking for immediately, without having to change apps. And with a touch we can go back to scrolling through the photos on the social network. The great thing is that this operation can also be done while we are watching a video. If there is something that interests us, just call up “Circle and search”. At that point the video pauses and you can take a look at the results proposed by the AI. Then you can immediately resume viewing.
“Circle and search” is not just an image search. For example, you can highlight a food present in a photo and you can ask the AI why it is so popular on social media. Or you can draw a circle around a plant and ask for advice on how to care for it. At that point Google’s AI will provide textual responses using its generative capabilities: it will not return a list of sites, but information written in natural language.
Artificial intelligence Perplexity AI, the intelligent search engine that challenges Google by Gabriella Rocco 13 January 2024
In short, the new AI does not forget the text, on the contrary. It puts it at the center of the research. Another example shown by Cathy Edwards, VP Google Searchduring Galaxy Unpacked, concerns the information contained in the messages we receive.
Edwards showed a message from a person suggesting dinner at a restaurant in Covent Garden, London, “that serves fried chicken with sriracha sauce”. The person who sent the message, however, did not mention the name of the venue. By turning on “Circle and Search,” Edwards was able to point out what little information was available. And based on the latter, the AI was able to find the restaurant and provide directions to reach it. With a touch she then returned to the conversation, again without the need to switch between apps. And above all without having to copy and paste the words of the message received on Google.
The surprising thing is that you can also highlight an “unconventional” text: a writing, for example, present in a video. A text of this type cannot be “pressed” and copied as you would the content of a web page, a post on social media or a document saved on Google Docs. Yet “Circle and Search” does just this: it uses generative AI to return detailed information on a phrase that appears in a video.
Also ChatGpt (OpenAI) e Copilot (Microsoft), Google’s two main competitors in terms of generative artificial intelligence, can analyze images. But this operation is still laborious: you have to photograph an object – or take a screenshot of a video – and then you have to upload the image online. Finally, a question is needed – which the user must necessarily type or dictate verbally – to set the AI in motion.