Google today announced a series of updates and new features at the Google I/O 2023, including advancements in Artificial Intelligence (AI) and new tools for users.
Help me Write in Gmail
Sundar Pichai, CEO of Google, unveiled a new feature in Gmail called “Help me write,” which allows users quickly draft smart emails. The feature, as shown in the demo by Google, can help in cases like asking for a refund from an airline. Just type the prompt and the AI-powered feature will automatically generate a draft based on the flight details in the previous email. Additionally, Google is also adding a Refine option that lets users refine their messages.
ALSO READ l Google I/O 2023 Live Updates: Google to bring generative AI to Search
This function provides several options, such as formalizing the wording, adding more details, shortening the message, or using the “I’m Feeling Lucky” option. Essentially, it allows users to tweak the language of their message to better suit their intended tone or meaning. The feature will roll out as part of Google Workspace updates.
Immersive view in Maps
Google Maps is also getting an update with “Immersive View” for routes, which uses AI visuals to provide preview renders of predicted traffic and weather. Immersive View for routes, a new feature for Maps will let users visualise their route before traveling. The technology fuses billions of Street View and aerial images together to create a digital model of the world, showing bike lanes, sidewalks, intersections and parking. Air quality information and traffic simulation are also included. The feature will be rolled out in 15 cities worldwide in the coming months. Google is also offering new APIs, including Aerial View and Photorealistic 3D Tiles, for developers to create customised and immersive experiences for their audiences.
Magic Editor
The company is working on a new feature called “Magic Editor” for Google Photos. The feature uses semantic understanding and generative AI to help users transform photos into works of art. It is a new experimental editing tool that uses generative AI to enhance users’ photos. With this feature, complex edits can be made without pro-level editing skills, such as repositioning the subject or changing the sky. Magic Editor also allows users to create new content to fill gaps and blend seamlessly into the photo. This new feature offers users more control over the final look and feel of their photos.
PaLM 2
As speculated, Google also unveiled its next generation language model- PaLM 2, at the event. The PaLM successor features improved multilingual, reasoning, and coding capabilities.
Google says PaLM 2 has been heavily trained on multilingual text and can now understand, generate, and translate nuanced text across more than 100 languages, including idioms, poems, and riddles.
PaLM 2’s dataset includes scientific papers and web pages with mathematical expressions, which has improved its capabilities in logic, common sense reasoning, and mathematics. The model has been pre-trained on publicly available source code datasets, enabling it to generate specialised code in popular programming languages like Python and JavaScript, as well as in languages like Prolog, Fortran, and Verilog.
ALSO READ l Watch: Google ad shows off Pixel Fold from every perceivable angle ahead of I/O 2023 launch
Google is also bringing together its Brain and DeepMind research teams to create the foundation models for AI, and is already working on its next model, Gemini, which will be “multimodal and highly efficient at tool and API integrations.”
Google Bard gets updates
The company also introduced updates for its AI-powered chatbot, Bard. Bard now knows 20 programming languages, has a dark mode, and can prompt users with images from Google search and their knowledge graph. Bard can also tap into services from other providers, including Instacart, Adobe, and Indeed.
Google has announced that its language AI model, Bard, will be available in over 180 countries and territories. The move comes after the model’s successful deployment in the US and UK, with the company also announcing it will support 40 languages soon. Additionally, Bard will soon have more visual responses to requests, with results accompanied by helpful visuals in addition to text. Users will also be able to upload images and include them in prompts, such as asking for a funny caption to be written along. Google Lens will be integrated with Bard to enable this feature. Developers will also receive upgrades including source citations, a Dark theme and an “Export” button to enable the exporting and running of code with Replit. Bard will also be integrated with popular Google apps such as Gmail, Drive and Maps, with more third-party services to be added in the coming months.
