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.
AI Video Tech Startup Moonvalley Raises $53 MLN
Nearly a month after Moonvalley, a Los Angeles-based startup developing AI tools for video creation, announced it had secured $43 million in fresh funding, the company has raised even more, according to a filing with the U.S. Securities and Exchange Commission.
The document, filed Thursday, reveals that Moonvalley has raised a total of around $53 million so far from a group of 14 undisclosed investors.
Moonvalley says it’s working with partners to manage licensing arrangements and compile videos into datasets the company then purchases. This approach is similar to that of companies like Pika and Adobe, the latter of which acquires training content from creators via its Adobe Stock platform.
OpenAI Launches Codex in ChatGPT
OpenAI announced on Friday a research preview of Codex, the company’s most capable AI coding agent to date.
Codex is powered by codex-1, a version of OpenAI’s O3 AI model, optimized specifically for software engineering tasks. OpenAI says codex-1 generates cleaner code than O3, follows instructions more precisely, and repeatedly tests its own code until it achieves successful results.
Codex runs in a protected virtual cloud environment. When connected to GitHub, Codex can preload your code repositories. OpenAI says the AI coding agent can take between 1 to 30 minutes to write simple features, fix bugs, answer questions about your codebase, and run tests, among other tasks.
OpenAI adds that Codex can handle multiple software engineering tasks simultaneously and doesn’t restrict users from accessing their computers or browsers while it’s running.
Google Rolls Out New AI and Accessibility Features for Android and Chrome
Google announced on Thursday a new AI-powered and accessibility features for Android and Chrome. Notably, the TalkBack screen reader on Android now lets users ask Gemini about screen content and images.
Last year, Google added Gemini capabilities to TalkBack to allow blind or low-vision users to access AI-generated image descriptions, even when alt text was missing. Now, users can ask questions and receive answers about their images.
Google also announced updates to Expressive Captions, Android’s real-time captioning tool. It now uses AI to capture not just what someone says, but how they say it.
Stability AI Launches On-Device Audio Generation Model
London-based AI startup Stability AI has launched Stable Audio Open Small, a stereo audio generation model that it claims is the fastest on the market, and efficient enough to run directly on smartphones.
The model was developed in collaboration with Arm, the chipmaker behind many processors in mobile devices. While AI apps like Suno and Udio can generate audio, most rely on cloud processing and can’t work offline.
Stability says the training set for Stable Audio Open Small consists entirely of royalty-free audio from Free Music Archive and Freesound, unlike Suno and Udio, which reportedly use copyrighted content, raising intellectual property concerns.
The model contains 341 million parameters, optimized for Arm processors. It generates up to 11 seconds of audio in less than 8 seconds on a smartphone, creating short samples and sound effects like drum beats and instrumentals.
DeepMind Unveils AI Tool for Math and Science Problem Solving
DeepMind, Google’s AI research lab, has developed a new AI system capable of solving problems using automatically classifiable solutions.
Called AlphaEvolve, the system may improve infrastructure Google uses to train its AI models. DeepMind says it’s building a user interface to interact with AlphaEvolve and plans to launch an early access program for select academic partners before broader release.
Google’s Gemini Can Now Analyze GitHub Projects More Easily
Gemini, Google’s AI chatbot, can now connect to GitHub, but only for users subscribed to the $20/month Gemini Advanced plan.
As of Wednesday, Gemini Advanced users can import public or private GitHub repositories directly into Gemini to generate, explain, and debug code, and more. Users can link GitHub to Gemini by clicking the “+” icon in the command bar, selecting “Import Code,” and pasting a GitHub link.
Tensor9 Helps Vendors Deploy Software Anywhere Using Digital Twins
Companies want access to new software and AI tools but are reluctant to expose sensitive data to third-party SaaS providers. Tensor9 aims to help software vendors reach more enterprise clients by enabling deployments directly within the customer’s own tech stack.
Tensor9 converts a vendor’s source code into the proper format for the client’s infrastructure, then creates a digital twin or miniaturized model of the deployed software, allowing clients to monitor performance remotely.
CEO Michael Tin Bao told TechCrunch that Tensor9’s ability to deploy software anywhere, including cloud and on-prem, and offer remote monitoring through digital twins, sets it apart from other vendors like Octopus Deploy and Nuon.
TikTok Launches “TikTok AI Alive” to Turn Photos into Videos
TikTok launched on Tuesday its first AI-powered photo-to-video feature, called TikTok AI Alive. It allows users to turn static images into animated videos for TikTok Stories.
Accessible via the Stories camera, AI Alive uses AI to generate short videos with “kinetic, atmospheric, and creative effects,” according to TikTok.
This launch follows TikTok’s earlier rollout of its AI image generator. While Instagram and Snapchat offer AI tools to turn text into images, TikTok is taking a step further by enabling users to create videos from photos. Snapchat, notably, has also announced a similar AI video tool coming soon.
Google to Bring Gemini AI to Android Auto
Google announced plans to integrate Gemini, its generative AI assistant, into all Android Auto-supported cars in the coming months, as revealed during the Android Show ahead of I/O 2025.
The company says that adding Gemini to Android Auto, and eventually to built-in car systems, will make driving “more productive and enjoyable,” according to a blog post.
Patrick Brady, Google’s VP of Android for Cars, said in a press briefing that this represents one of the biggest shifts in in-car experiences in a long time.
Gemini will serve as a smarter voice assistant, allowing drivers (or passengers) to send messages, play music, and perform tasks without rigid voice commands, thanks to its natural language capabilities.
Google Launches AI Futures Fund to Support Emerging AI Startups
Google has announced the AI Futures Fund, a new initiative to invest in startups building with the latest AI tools from Google DeepMind.
The fund will support startups from early-stage to growth and offer varied levels of assistance, including early access to DeepMind models, collaboration with Google experts, and Google Cloud credits. Select startups may also receive direct investment from Google.
New AI Tool Predicts Biological Age by Analyzing Your Face
A new AI tool developed by researchers at Mass General Brigham and Harvard Medical School can predict a person’s biological age, and survival after a cancer diagnosis, just by analyzing a facial photo.
Chronological age is based on date of birth, but biological age reflects physiological health and expected lifespan. Dr. Hugo Aerts, lead co-author and AI in Medicine Program Director at Harvard, explains: “Biological age depends on lifestyle, genetics, and health. We hypothesized that a person’s appearance might reflect that.”
The team trained their deep learning model, FaceAge, on over 58,000 images of healthy individuals and 6,000 images of cancer patients with known outcomes.
The algorithm showed that cancer patients’ FaceAge was, on average, five years older than their actual age, and looking older correlated with worse outcomes across cancer types.
