The artificial intelligence (AI) realm witnesses rapid developments around the world, reshaping the world of digital transformation and accelerating the pace of AI adoption. In this weekly newsletter, FollowICT is highlighting the most important news and advancements in AI around the globe.
Amazon Tests New AI Shopping Agent
Amazon has started testing a new AI shopping agent, branded as “Buy for Me,” with a subset of users, the company announced in a blog post on Thursday.
If Amazon does not sell a product that users are searching for, the “Buy for Me” feature will suggest products available on other sites. Users can then choose one of these products and request to purchase it without leaving the Amazon shopping app.
Amazon is the latest company to unveil a smart shopping agent, joining firms like OpenAI, Google, and Perplexity, all of which have introduced similar agents that can browse websites and assist users in making purchases. Amazon is already the preferred platform for most people to buy anything they want online, but the “Buy for Me” feature could help Amazon attract more e-commerce business compared to what it currently achieves.
Microsoft Brings Copilot Vision to Windows and Mobile
Microsoft is rolling out the Copilot Vision feature to Windows and mobile platforms, extending beyond the web. Copilot Vision was a key component of Microsoft’s Copilot redesign last year, but until now, it has been limited to Edge web pages to help users explore what they see.
Copilot Vision will now be able to analyze live video from a mobile phone camera, allowing the AI assistant to provide tips for keeping plants healthy or decorating your home. Microsoft is launching the Copilot Vision app within the Copilot app for iOS and Android, with a similar feature coming soon for Windows.
Competition Between OpenAI and Anthropic for College Students
Leading AI labs OpenAI and Anthropic have announced major initiatives in higher education. It’s the ongoing competition we’re all familiar with: last week, Anthropic dropped its announcement at 8 AM on Wednesday, followed closely by OpenAI with a similar news release at 8 AM on Thursday.
For Anthropic, last week’s announcement marked its first major academic push. The company launched “Claude for Education,” a version of its chatbot tailored for universities. It also announced partnerships with Northeastern University, the London School of Economics, Champlain College, and Internet2, which builds technical infrastructure for universities, as well as Instruction (the developer of the Canvas program), aiming to enhance “equitable access to tools that support universities in integrating AI.”
At the core of Anthropic’s educational offerings is “Learning Mode,” a new feature that changes how Claude interacts with students. Instead of just providing answers, the press release states that “Learning Mode” will use the Socratic method to guide students in problem-solving by asking questions like “How would you approach this?” or “What evidence supports your conclusion?”—with the goal of helping students “develop critical thinking skills” rather than just doing their homework for them.
While Anthropic has just entered the higher education space, OpenAI has been active in this area for nearly a year. The startup launched ChatGPT Edu in May 2024, a version of its chatbot designed for universities, along with multiple partnerships with colleges. Last month, the company formed the NextGenAI alliance, dedicating $50 million to accelerate AI research across 15 colleges. In February, OpenAI also partnered with California State University to provide ChatGPT Edu across all CSU campuses. Then, last Thursday, OpenAI announced that ChatGPT Plus (at $20 a month) would be free for all college students in the U.S. and Canada until May.
Adobe Launches AI-Generated Video Expansion Tool for Premiere Pro
Adobe is updating Premiere Pro with AI-powered features aimed at providing creators with faster and better video editing results. The release of Premiere Pro version 25.2 adds tools for locating, translating, and expanding video clips, moving from beta to general availability for all users.
The standout feature is Generative Extend, announced by Adobe in October as one of the first tools using the Firefly generative AI model for video. This feature enables users to extend clips by up to two seconds, offering more options for transitions or correcting unexpected movements without needing to reshoot footage. Generative Extend can now create clips in 4K quality, extending ambient background sound—up to ten seconds for just the sound, or two seconds when paired with video extension—without prolonging speech or music.
According to Adobe, the Generative Extend feature is completely free to use for a limited time, after which it will require users to spend generative Firefly credits. Creative Cloud subscriptions come with monthly credits ranging from 25 to 1,000 credits, depending on the plan. An additional Firefly subscription starts at $10, providing 2,000 credits monthly. Adobe has not specified how many credits the Generative Extend feature will ultimately consume, but it noted that “the price varies based on video format, frame rate, and resolution.”
OpenAI Plans to Release its 1st Open AI Language Model
OpenAI has announced its intention to release its first open language model since GPT-2 in the coming months, according to a posting on the company’s website on Monday.
The model includes questions for developers and researchers, such as: “What would you like to see in an open model from OpenAI?” and “What open models have you used in the past?”
OpenAI wrote on its website: “We are excited to collaborate with developers, researchers, and the community as a whole to gather feedback and make this model as useful as possible.”
Microsoft to Allow Users to Customize AI Assistant
Microsoft, determined to maintain its position in the AI field, will soon allow consumers to customize the Copilot digital assistant according to their specific needs.
The chatbot will be able to capture user preferences—such as favorite foods, movies, or past vacations—and carry that information into future conversations. These features, set to roll out in the coming months, reflect efforts by competing chatbot companies to personalize their products.
AI Video Startup Valued at $3 bln in Funding Round
Runway AI Inc. has raised $308 million in a new funding round, more than doubling the company’s valuation—an indication of investor enthusiasm for startups building AI software capable of creating videos.
The deal, set to be announced on Thursday, raises Runway’s valuation to just over $3 billion, according to Bloomberg. The funding round was led by private equity firm General Atlantic, which closed late last year. Other investors include Nvidia and the Vision Fund 2 from SoftBank.
The Runway platform offers a range of AI-powered media tools, including video and image generation models. It faces stiff competition in video generation from companies like OpenAI and Google. However, the company has struggled to differentiate itself, having secured a deal with a major Hollywood studio and allocated millions to fund films using AI-generated footage.
This week, Runway launched Gen-4, a video generation model the company claims can create consistent characters, locations, and objects across scenes, maintain “cohesive world environments,” and regenerate elements from different perspectives and positions within scenes. With products like Gen-4 and a recently launched API for video models, Runway aims for $300 million in annual revenue this year.
GitHub Copilot Introduces New Limits and Charges for “Premium” AI Models
GitHub Copilot, Microsoft’s AI coding assistant, may soon become more expensive for some users.
GitHub announced on Friday the introduction of “premium requests” for GitHub Copilot, a new system that imposes speed limits when users switch to AI models that are not the base model for tasks like “agent” programming and editing multiple files. While GitHub Copilot subscribers can still perform an unlimited number of actions with the base model (OpenAI’s GPT-4o), there will now be caps on tasks and actions using newer models, such as Anthropic’s 3.7 Sonnet.
Customers on the Copilot Pro plan (priced at $20 per month) will receive 300 premium subscription requests monthly starting 5 May, according to a GitHub blog post. Users on the Copilot Business and Copilot Enterprise plans will receive 300 and 1,000 premium subscription requests respectively, starting 12 May and 19 May.
Customers subscribed to any of these plans can purchase additional premium requests at $0.04 each or upgrade to GitHub’s new Copilot Pro+ plan. The Copilot Pro+ plan starts at $39 per month and offers 1,500 premium requests and “access to the best models,” according to GitHub, including OpenAI’s GPT-4.5.