Hello readers,
Welcome to your weekly AI newsletter, where we summarize the top stories in artificial intelligence. This week saw several major developments that will shape the future of AI.
First, Adobe launched its much-anticipated generative AI tool Firefly commercially after months in beta. Creators can now access features like image generation and expansions in Photoshop and other Adobe apps.
Also making waves is Google's upgrade to its conversational AI chatbot Bard, with new capabilities to improve accuracy and user trust. This positions Bard as a potentially powerful personal assistant integrated with Google services.
Finally, Microsoft unveiled its vision for an everyday AI companion called Copilot to assist across work, life and play. The ambient assistant promises personalized help by unifying Microsoft's AI strengths.
With powerful new generative AI capabilities, the race heating up in AI assistants, and increasing democratization of these technologies, it’s an exciting time in AI. Read on for the details!
This newsletter is also available in Spanish and Catalan.
Adobe's generative AI Firefly is now commercially available
On September 13, 2023, Adobe announced the commercial launch of Firefly, its generative AI model, after 176 days in beta. Firefly is now available in Creative Cloud, Adobe Express and Adobe Experience Cloud. This means features like generative fill and generative expand in Photoshop are now available without having to install the beta. The announcement is important as Adobe also detailed how it will charge for using Firefly, through "generative credits", allowing more equitable access to this technology.
TikTok Introduces New AI Labeling to Boost Transparency
TikTok has rolled out new tools to help creators label AI-generated content on the platform. Creators can now easily tag videos made with AI, and TikTok will begin testing automatic AI detection to label some content. The move aims to boost transparency around AI as its use proliferates. TikTok will also rename effects using AI and is partnering to develop better AI detection.
Google Upgrades Bard with New Features for Better Answers
Google released an upgraded version of its AI chatbot Bard with new capabilities to improve the accuracy of its responses. Users can now double-check Bard's answers against Google search results, collaborate with others on Bard conversations, and connect Bard to their Google apps like Gmail, Docs, and Maps. These features address key challenges with AI chatbots around trust and factual correctness. The integration with Google services also shows Bard's potential as a personalized digital assistant. As the race heats up between Google and competitors like Microsoft and Meta in AI chatbots, this Bard update aims to make Google's offering more useful while building user trust through transparency.
Teaching Robots New Tricks Overnight: Toyota Researchers Pioneer Large Behavior Models
In an exciting advance, Toyota Research Institute (TRI) has developed a new method to rapidly teach robots skills using a technique called Large Behavior Models. This method allows robots to learn complex manual tasks overnight through AI, instead of needing millions of training examples. By combining robot learning approaches with diffusion models similar to DALL-E, TRI has already trained robots in 60 skills with high success rates. The researchers believe this breakthrough could lead to more capable robots that can adapt to diverse settings and interact with novel objects, moving us closer to truly "general purpose" robot intelligence.
Google and the DoD Develop AI-Powered Microscope to Detect Cancer
Google and the U.S. Department of Defense announced they have collaborated to build an augmented reality microscope (ARM) powered by artificial intelligence. The ARM is designed to help pathologists identify cancer more accurately and efficiently. When a slide is viewed under the microscope, AI outlines the cancer in green and provides a heat map showing the extent of the disease. Initial research on breast cancer detection has been promising.
GitHub Expands Access to AI-Powered Copilot Chat to Individual Users
GitHub announced that it is expanding access to Copilot Chat, its AI-powered programming assistant, to all individual GitHub Copilot subscribers. Copilot Chat was previously only available to business customers. The chatbot is integrated directly into Visual Studio Code and provides tailored coding suggestions and assistance based on the code developers are working on. Making Copilot Chat available to individual developers is an important step in democratizing AI and enabling "natural language as a new universal programming language," according to GitHub's VP of Product.
Microsoft unveils Copilot - the everyday AI assistant
Microsoft announced the launch of Copilot, an AI-powered digital assistant designed to help users across work, life, and play. Copilot brings together AI capabilities across Microsoft's products and services, including Windows, Microsoft 365, Teams, Outlook, Bing and more. The assistant can understand context and user history to provide personalized help, take actions, and boost productivity. Microsoft is positioning Copilot as the next evolution of AI, moving beyond just answering questions to actively assisting with tasks.
DALL-E 3 Coming in October: A Leap Forward in Imagination
OpenAI has announced that DALL-E 3, the latest iteration of its revolutionary AI image generator, will be available in October 2023. DALL-E 3 represents a significant improvement in image quality and fidelity over previous versions. Built natively on top of ChatGPT, DALL-E 3 can now take simple text prompts and turn them into stunningly realistic and creative images. With nuanced understanding of language and world knowledge, DALL-E 3 promises to unlock new levels of imagination and drastically change how we create and interact with visual media.
That wraps up this week's AI news! If you found this informative, don't forget to share it with friends and colleagues. And be sure to subscribe to get next week's news straight to your inbox. Thanks for reading!