Google releases AI image watermarking tool and opens "big model store"

Google is working to position itself as a platform that gives customers the freedom to choose the AI ​​model that best suits their needs, whether from Google or its partners. Google Cloud users now have access to more than 100 powerful artificial intelligence models and tools

Google has announced a new tool called SynthID for embedding watermarks to indicate that images were created by artificial intelligence. This feature is technically supported by artificial intelligence laboratory Google DeepMind and includes pixel-level watermarks, making images difficult to tamper with

Google Cloud CEO Thomas Kurian said it has enough resources to meet customer needs.

In a bid to compete in the lucrative cloud software market, Google Cloud CEO Thomas Kurian introduced a slew of new products at Google's annual cloud conference in San Francisco and announced an offer that can help draft emails. Or a tool to summarize cloud documents

These new products will intensify competition between Google and Microsoft as the second largest cloud provider and its support for OpenAI. Meanwhile, Google CEO Sundar Pichai wants to reduce reliance on search engines as a primary source of revenue

Rewritten content: Users of Google Cloud services will be able to access Meta's large language model Llama 2 and artificial intelligence startup Anthropic's Claude 2 chatbot to customize their own applications and services using enterprise data. Google is working hard to position itself as a platform where customers have the freedom to choose the AI ​​model that best meets their needs, whether from Google or its partners. Google Cloud users now have access to more than 100 powerful AI models and tools, the company said

Google said that in addition to adding new artificial intelligence models to its cloud services, it is also improving its own artificial intelligence models and tools. The company says its large language model, PaLM 2, now available in 38 languages, can better analyze longer documents such as research papers, books and legal abstracts. The company's text-to-image app Imagen will offer updated features such as higher-quality images and style adjustments

Google announced that it will make its Duet AI product more widely available to users of its Workspace productivity suite, with plans to make it available to the public later this year. Users can click to generate an AI assistant that responds to prompts to help create content on apps like Google Docs, Sheets, and Slides. Google says Duet AI can take notes during video calls, send meeting summaries and translate subtitles in 18 languages. With a new feature called Join Me, users can schedule the tool to attend meetings on their behalf, deliver messages and create event recaps

When you get lost in a meeting, Google can help you fill in what you missed

Google also announced new partnerships with companies including General Electric and Fox Sports to enable customers to take advantage of artificial intelligence technology. These technologies include creating customized recipes or watching replays of sporting events

Google announced a deepening partnership with NVIDIA. Google said it will expand its cloud services to make greater use of Nvidia's chips and products to speed up the training of large language models. Google also said it can use Nvidia's H100 chip and allow customers to use the latest version of Nvidia's supercomputer

The company noted that some industry-specific models are growing in popularity, such as Med-PaLM 2, an artificial intelligence model suitable for medical settings. Google currently has partnerships with healthcare companies such as Bayer Pharmaceuticals. In addition, the company has developed the Sec-PaLM 2 model specifically designed for network security, which has been adopted by some network security providers

As technology companies race to improve their artificial intelligence products, AI-generated images are becoming increasingly difficult to distinguish from real images. Google has announced a new tool called SynthID for embedding watermarks to indicate that images were created by artificial intelligence. This feature is powered by technology from Google DeepMind, the artificial intelligence lab owned by Google, and includes pixel-level watermarks that make images difficult to tamper with. Microsoft has formed an alliance with technology and media companies to jointly develop standards for watermarking AI images and is working on new ways to track AI images

The above is the detailed content of Google releases AI image watermarking tool and opens "big model store". For more information, please follow other related articles on the PHP Chinese website!