Google organised its annual developer’s conference, Google I/O 2023 last night. As with each year, the main focus of I/O was Google’s software, although there were a few hardware launches as well. For software though, 2023 is clearly going to be a year for Google’s AI projects if there ever was one.

Google has often teased audiences about its upcoming AI-based projects and features in previous editions of the I/O, albeit with a careful and measured approach. However, this time around, Google has been forced to go all out with its AI offerings.

Here are some of the biggest AI-related announcements that Google made at I/O 2023.

Bard opening up to all
Google has revealed that its competitor to ChatGPT, known as Google Bard, will now be accessible in more than 180 countries, with the waitlist no longer in effect. During the Google I/O 2023 event on May 10, Google shared the news that Bard would soon become available to additional countries and regions.

Simultaneously, Google presented forthcoming enhancements for Gmail, including an AI-powered tool designed to assist users in composing emails based on specific prompts. This particular feature is anticipated to be launched within this year.

Furthermore, Google has announced that Bard will soon offer support for languages such as Japanese and Korean. The software giant has also made a commitment to adding support for an additional 40 languages, though the specifics of those languages have not yet been disclosed.

A revamped Search and a refreshed Bard
To enhance user experience, Google Bard is incorporating more visual elements. For example, users now have the ability to request the AI to search for “must-see sights” in a specific region, and the AI, utilizing the power of Google Search and Knowledge Graph, will present relevant results accompanied by images.

This is a feature that ChatGPT does not possess. Additionally, Bard allows users to customize their viewing style, such as opting for a tile or list format.

Google is also making improvements to Google Lens, a tool that enables users to search for objects within images. With the latest enhancements, users can upload an image to Lens and request the AI to generate a humorous caption for it.

Furthermore, Google emphasizes that Bard’s citations will become more precise, aiding users in locating accurate sources of information. A new “export” feature will be introduced, allowing users to directly upload the generated text into Gmail or Google Docs for seamless integration.

Regarding collaborations and image-related advancements, Google is partnering with Adobe, the company behind Adobe Firefly, which enables creators to generate original images based on text prompts. This collaboration will enable users to create high-quality images effortlessly by providing text inputs such as “Create a birthday invitation for my unicorn-loving daughter.” The AI-powered tool will swiftly generate an image incorporating relevant elements based on the input.

PaLM 2: a new large-language mode
Google introduced PaLM 2, a new large-language model with impressive capabilities across various domains. PaLM 2 is capable of performing tasks such as coding, reasoning, multilingual translation, and natural language generation.

It has been trained on a vast collection of over 100 spoken languages, enabling it to excel in multilingual tasks that require understanding nuanced phrases, including idioms, poems, and riddles, which often involve ambiguous and figurative word meanings.

One notable aspect of PaLM 2’s training is its exposure to extensive math and science texts, encompassing scientific papers and mathematical expressions. This training has enhanced the model’s reasoning abilities, which is evident in its problem-solving capabilities, particularly in puzzle-solving scenarios.

Google emphasizes that PaLM 2 powers more than 25 Google products behind the scenes, showcasing its wide-ranging application within the company’s services. Additionally, Google recently introduced a specialized version of PaLM 2 called “Med-PaLM 2,” which focuses on medical knowledge. It has demonstrated remarkable performance, capable of answering medical exam questions at an “expert doctor level” with an impressive 85% accuracy rate.

AI meets Gmail – Google’s ‘Help Me Write’ feature
Google has introduced an enhanced email composition feature on Gmail that leverages AI and Google paML2 technology, enabling users to create emails rapidly. This feature, referred to as “Help Me Write,” eliminates the need for downloading ChatGPT-powered extensions.

During the Google I/O 2023 event, Google showcased a new pencil icon within Gmail, which allows users to utilize prompts when composing emails. For example, if a user receives an email regarding cancelled flights, they can click on the pencil icon and type “Ask for a full refund.” A smart reply will be generated based on this input.

Google emphasizes that the suggested reply is provided as a reference, and users have the freedom to modify or adjust the text according to their personal style. If users are dissatisfied with the initial suggestion, they have options such as “recreate” and “elaborate” to further refine their email response.

Read all the Latest News, Trending News, Cricket News, Bollywood News,
India News and Entertainment News here. Follow us on Facebook, Twitter and Instagram.