The technology giant Google held its traditional Google I/O developer conference on Wednesday in its return to in-person events. During the presentation the generative Artificial Intelligence Bard, the Android 14 operating system and the expected presentation of its folding phone Pixel Fold dominated the announcements.
Sundar Pichai, CEO of the company and Alphabet, was the main speaker at the event that took place in the auditorium that Google has at its headquarters in Mountain View, California. In his presentation, he admitted that he seeks to make Bard a tool of help for everyone.
MSundar Pichai, CEO of Google and Alphabet. Photo: AFP.
In the search to fulfill that purpose, Google opened the smart and free tool to 180 countries, initially for the English, Japanese and Korean languages, which it launched almost two months ago and that until now could only be accessed by invitation or by waiting list.
Sissie Hsiao, vice president of Google and general manager of Google Assistant and Bard, said that "very soon" Bard will be available in "the top 40 languages" of the world.
Another new feature is that users will be able to export responses obtained by Bard to Gmail and Google Docs. "Because people often ask Bard for help in composing emails and documents," Hsiao said.
One of the novelties most applauded by the multitudinous audience was that the chatbot can also be used on a black background and not only white, as until now.
How is Bard, Google's ChatGPT
During the Google I/O, the main developments of the technological giant in artificial intelligence, operating systems and cellular systems were presented. Photo: REUTERS/Jeffrey Dastin
As Microsoft announced last week, Google's chatbot will "soon" give multimodal answers, that is, it will not only respond with text, but also with images generated thanks to this generative artificial intelligence in the style of Midjourney.
This new function will operate in two ways: on the one hand, it will incorporate images in the answers, but on the other hand it will be the users themselves who will be able to ask Bard questions with photos.
An example of the latter case is that a user can upload a photo -with Google Lens image recognition- of a drawer full of art supplies and ask the chatbot to prepare a list of activities that can be practiced with children with the materials in front of them.
"In the coming months, we will integrate Adobe Firefly into Bard. That way, people will be able to generate new images," he noted.
Last week Microsoft integrated the DALL-E 2 technology of Open AI (creators of ChatGPT and DALL-E) to allow its users to create images using a text as well.
Starting next week, there will be more transparency in the source of the information in Bard's response, the company promised.
"Starting next week, we'll make the code sections even more accurate by displaying the specific code blocks that are obtained alongside any relevant licensing information. And this will also apply to the quotation of narrative content from across the web," the company announced.
In addition, Bard's extensions will apply to many of the apps and services people already use in Google Maps, Google Sheets, Gmail, Docs, and more.
Google's popular Internet search engine will also receive new capabilities, obviously, powered by generative artificial intelligence. This points to much more accurate and more frequently updated results which, according to the tech company, is estimated at 1.8 trillion for now.
However, the company did not give a date for this last step.
PaLM 2, the revolutionary language model
PaLM 2 powers the artificial intelligence Bard, rival of ChatGPT. Photo: Bloomberg.
On the other hand, the Google I/O event had another big announcement with the launch of PaLM 2, Google's new language model (LLM), which is also behind Bard.
This is an update that will allow many improvements, such as skills in mathematics, logic and reasoning, and more recently, capabilities to write programming code and even find errors and correct them (debugging), according to Pichai.
PaLM 2 includes more than 100 languages and had training in 20 programming languages.
Not only will it have programming skills, it also promises to provide healthcare for radiologists, as PaLM2 provides the ability to recognize X-rays.
Google introduced a new language model for training artificial intelligence tools such as chatbots, known as PaLM 2. Photo: Bloomberg.
Google thus continues its race with Microsoft, a company that last week opened to everyone its free artificial intelligence tool, Bing, which has the support of OpenIA, developer of ChatGPT and DALL-E, capable of generating texts and images.
A smarter image editor
Google's new Magic Editor photo editor is powered by generative artificial intelligence (AI) and can totally transform photos without professional tools, the company announced Wednesday at its annual Google I/O 2023 developer event.
Sundar Pichai, CEO of Google and Alphabet, said that, with the help of Magic Editor's generative AI, a "new experimental editing experience," Google will help users make complex edits to photos with a simple click and shared a couple of examples with the audience who followed the event in person and online.
In the first, he showed a photo of a person in front of a waterfall and showed how using this editor you could move the person to make the waterfall look better, erase the strap of a bag and make the sky less cloudy.
In another example, Magic Editor was able to move a child on a bench closer to the center of the photo, generating a "new" part of the bench he was sitting on.
Google and Apple, united
During the conference there was also the announcement of a new initiative between Google and Apple. Both giants proposed a new technical specification that alerts people who are being tracked through small devices initially sold to find their keys, luggage or other personal items.
This is the case of Apple's AirTags, coin-sized devices equipped with Bluetooth technology, which are popular with travelers but have also been implicated in cases of harassment.
Connected to a mobile application, they allow to follow in real time the geographical position of the article to which they are attached but also of the people who transport it, sometimes without their knowledge.
iPhones, for example, already notify their users if they detect a nearby AirTag (or Apple's wireless headphones) that doesn't belong to them while on the move.
The proposal of the two technological giants should serve to make Bluetooth tracking devices compatible with the detection and alert systems of the iOS and Android operating systems.
Samsung, Tile and other makers of similar devices "have voiced their support" for this new industry standard, according to the joint statement from Apple and Google.
"Bluetooth devices have immense practical benefits for users, but they also have the potential to be used to track people without their knowledge, (a problem) that the entire industry has to solve," Dave Burke, vice president in charge of Android engineering, said in the statement.
In the United States, several women took legal action against Apple in 2022, after being tracked through an AirTag and harassed.
In June of last year, a 26-year-old man was killed by his girlfriend, who suspected him of cheating and tracked his location through an AirTag in the state of Indiana, according to court documents.
And in February, Robert Reeves, a spokesman for the Irving, Texas, police admitted to the AFP news agency that the city's police station had already handled several cases related to the famous Apple accessory, in which the victim and his follower knew each other.
Pixel Fold, the first of Google's folding
Rick Osterloh introduced the Pixel Fold. Photo: AP.
Then came the main course with the presentation of the first of the Pixel Fold, Google's folding screen phones that will compete with the Galaxy Z Fold 5 that Samsung may present in August.
This version transforms from a traditional 5.8-inch phone to a 7.6-inch tablet thanks to its flexible screen and hinge, features a 120Hz refresh rate and Gorilla Glass Victus protection.
The Pixel Fold is also premium inside: it has the same Tensor 2 processor as the Pixel 7 and 7 Pro (as well as in the Pixel 7A and Pixel Tablet, which were also announced during the event) along with 12GB of RAM and configurations ranging from 256 or 512GB of storage.
On its autonomy, Google accuses "more than 24 hours" of battery life of 4800 mAh, which is divided into two cells, one in each half.
The price of the Pixel Fold. Photo: AP.
In addition, Google's native applications are more responsive and adapted for this format, such as dual-screen translator or video playback.
In total, Pixel Fold has five cameras, two of 8.3 megapixels, one on the outside for selfies and another above the internal screen for video calls. On the rear it has a module of three sensors: a main one with 48-megapixel optical stabilization, a 10.8-megapixel ultra-wide camera and a 10.8-megapixel camera. -Telephoto 5x megapixel.
See also