A birds-eye view of the top AI news for May, 2023.
Adobe Integrates Firefly Generative AI into Photoshop
Adobe has integrated generative AI capabilities into Photoshop. Users can now access Adobe Firefly, Adobe’s family of generative models, directly from within Photoshop. The graphics editor allows users to create images with Generative Fill using natural language prompts, add content, remove and replace parts, or extend the edges of an image.
Adobe is also actively working on a compensation model for Adobe Stock contributors, stating that Firefly’s initial model was trained “on a unique dataset that includes Adobe Stock images, openly licensed content and other public domain content without copyright restrictions.”
In addition to Generative Fill, Adobe introduced new tools to accelerate everyday creative workflows in Photoshop, including adjustment presets, and a contextual task bar.
Adobe Firefly entered the generative AI space six weeks ago. According to the company, Firefly has become one of the most successful beta launches in its history, with 100 million assets have been generated by its beta users to date.
OpenAI’s CEO Sam Altman Suggests Creation of AI Policy Agency
Sam Altman, CEO of OpenAI, took the stage at a Senate hearing on May 16 to urge US lawmakers to regulate the AI industry. He emphasized the importance of licensing and independent oversight of AI companies to mitigate the risks posed by the technology’s rapid growth. Altman suggested the creation of “a new agency that licenses any effort above a certain threshold of capabilities.” Joining Altman at the hearing were Christina Montgomery, IBM’s chief privacy and trust officer, and Gary Marcus, a former New York University professor.
Some experts were critical of Sam Altman’s perspective, arguing that more near-term risks need to be mitigated rather than potential future harms, according to Axios. Others say the call for AI regulation is a way for AI leaders, including OpenAI, to limit competition, as small startups are likely to face barriers to entering the industry.
Later this month, Sam Altman came to London to talk about AI at University College London, where he faced a group of protesters calling on OpenAI to stop building AGI, according to The Verge.
Wrap of Google’s I/O 2023
At its annual developer event, Google I/O, the company unveiled a number of AI advancements. Here are some of the more interesting announcements from the keynote.
- AI automation has already been working behind the scenes in Google Workspace, with 180 billion AI-powered jobs performed last year, including grammar, spell checking, smart composing, and replying in Gmail, Docs, and other applications within the suite. Now, Google has announced the deep integration of AI into Google Workspace, introducing Duet AI tools that enable workflow automation. For example, the AI assistant will soon be able to consider the context of the thread when drafting responses in Gmail or generate attractive images for a presentation in Slides.
- In the realm of photography, Google Photo, powered by the generative AI Magic Editor, promises to enhance photos with impressive capabilities. Google plans to give select Pixel phones early access to the feature later this year.
- Incorporating Google’s latest advances in large language models (LLM), including the versatile PaLM 2.0, into the Bard chatbot takes the AI interaction experience to the next level with a deeper understanding of diverse forms of media — Bard will soon be able to support images in prompts, and by the end of the year, the chatbot will be available in the 40 most spoken languages. Today, users in over 180 countries can chat with Bard in English, Korean, and Japanese.
- In addition, Google revealed its work on Gemini, a next-generation large language model that aims to rival Microsoft-backed GPT-4 by OpenAI (which released the ChatGPT iOS app this month, with an Android app coming soon, and reached half a million installs in less than a week).
- Notably, Android had AI as one of its main focuses, underscoring Google’s commitment to AI. The upcoming version of Android will place a strong emphasis on integrating AI technologies to enhance user experience.
Shutterstock Acquires GIPHY from Meta for $53 Million
Shutterstock, a global stock imagery platform, has announced the acquisition of Giphy from Meta for $53 million.
Giphy, known for its vast library of animated GIFs and stickers, draws more than 1.3 billion daily searches and drives more than 15 billion daily media impressions, with estimated annual revenue of $26.6 million, according to Growjo. The platform is expected to be integrated into Shutterstock, giving users access to a wider range of visual content for their creative projects.
The acquisition of Giphy allows Shutterstock to leverage the popularity and broad reach of GIFs across social media, messaging apps, and digital communications to enhance its own offerings and strengthen its position as a go-to platform for creative professionals and marketers.
We covered the history of GIPHY in our previous posts — The History of Emoji, GIFs, Stickers, and Avatars. How Nonverbal Communication Went Digital and How Stickers, Avatars, and GIFs Became a Commodity.
Amazon and Snap Embrace AI for Innovative Ad Products
Amazon is taking steps to expand its advertising business by harnessing the power of AI. According to The Information, the company is forming a team to develop AI tools that will generate photos and videos for merchants to use in their ad campaigns. These efforts were later spotted by Bloomberg in Amazon’s job listings, demonstrating the company’s ambitions to develop ChatGPT-like interactive conversational experiences. While Amazon’s current ad business is primarily focused on increasing visibility in search results, the introduction of AI-generated content represents an opportunity to diversify its offerings and drive ad revenue.
Snap, the parent company of Snapchat, is also making strides in the advertising space with the help of AI. One notable development is the integration of sponsored links into conversations on its My AI chatbot. This allows users to receive sponsored links that are relevant to their ongoing conversations, such as recommendations for restaurants, hotels, or video games.