Welcome to AgiYes‘s AI News Weekly!
We deliver the most impactful breakthroughs, cutting-edge research, and real-world innovations from across the globe.
We’ve curated a list of the 10 most significant AI developments this week to keep you informed. Let’s dive in!
Top 10 AI News from August 9–15, 2025
1. Anthropic Rolls Out a New Learning Mode for Claude
Anthropic has announced a new “Learning Mode” for its popular AI chatbot, Claude, and the code-centric version, Claude Code. The new feature aims to transform the chatbot from a simple “answer dispenser” into an interactive teaching partner.
At the heart of this update is the Socratic method, which guides users toward understanding by prompting them with exploratory questions instead of providing direct answers. This mode was initially a limited trial for the education sector in April and is now available to all users via a dropdown menu.
2. Google Launches Gemma 3 270M, a Compact and Efficient Open-Source AI Model
Google’s DeepMind AI research team recently unveiled Gemma 3 270M, a new open-source AI model. With just 270 million parameters, the model is significantly smaller than the large language models (LLMs) that typically have 70 billion or more parameters.
Despite its smaller size, Google’s goal is to provide a highly efficient model that developers can run offline directly on smartphones. Internal testing has already shown promising performance on the Pixel 9 Pro’s System on a Chip (SoC).
3. OpenAI Considers Introducing Ads to ChatGPT
According to a report from The Verge, OpenAI is exploring various new revenue streams, including the potential for introducing ads to ChatGPT. In an interview, ChatGPT’s head, Nick Turley, stated he “wouldn’t rule it out entirely” but emphasized that any ad integration would need to be “very careful and tasteful” to avoid negatively impacting the user experience.
ChatGPT’s user base has now surpassed 700 million, with approximately 20 million paid subscribers. Turley views the large free user base not as a burden but as a significant opportunity to create a differentiated paid product.
4. Google’s New AI Tool Flight Deals Simplifies Trip Planning
Google has announced the launch of “Flight Deals,” a new AI-powered tool for its Google Flights service that helps travelers find discounted flights using natural language queries.
The company says Flight Deals is designed for “flexible travelers” who can enter a free-form description of their trip plans into the search bar. The tool then uses a customized version of the Gemini 2.5 model to analyze the user’s request and match it with real-time flight data.
5. Google Photos to Help Users Identify AI-Edited Images
As AI capabilities advance, it’s becoming increasingly difficult to distinguish between real and fake images. Recognizing growing public concern about AI-generated content, Google is launching a new feature to help users verify the authenticity of photos and videos.
The upcoming Google Photos feature, reportedly named “How Was This Made,” aims to help users quickly determine whether a photo or video was captured naturally, modified with editing tools, or generated entirely by AI.
6. Mistral AI Unveils Major Update with Medium 3.1
Mistral AI has officially released its latest AI model, Mistral Medium 3.1, marking a significant step forward in the company’s natural language processing capabilities. The new version features comprehensive upgrades in performance, tone, and web search intelligence. Users can now access the new model directly through Mistral AI’s “Le Chat” platform or via its API (with the model name “mistral-medium-2508”).
One of the key highlights of Mistral Medium 3.1 is its enhanced performance, particularly in core reasoning and coding. The model shows significant improvements in accuracy and speed across tasks like code generation, STEM reasoning, and cross-domain Q&A.
7. Midjourney Adds HD Video Generation for Standard Subscribers
Midjourney has rolled out an exciting set of new features, significantly enhancing the creative experience for its standard subscribers. Now, standard users can generate high-definition videos with greater creative flexibility and bulk production capabilities, opening up new possibilities for content creators.
The update also includes an optimized video moderation function to improve the accuracy and compliance of content. This ensures that videos created by users better align with community standards, giving creators greater confidence when sharing their work.
8. DeepSeek App Updates to Version 1.3.0
On August 14, the DeepSeek App released its 1.3.0 update for mobile app stores. The main highlight of this new version is the addition of a feature that allows users to generate and share images of their chat conversations.
This native feature eliminates the need for manual screenshots, providing a more convenient and visually appealing way for users to share interesting conversations. It’s designed to significantly improve the experience of sharing content on social media.
9. Kimi to Launch a Global PPT Generation Feature Powered by K2
Moonshot AI’s Kimi model is set to launch a global presentation (PPT) generation feature powered by its robust K2 model. This breakthrough has garnered significant attention from the industry.
Leveraging the Kimi K2 model’s superior performance in code generation, mathematical reasoning, and agent-based tasks, the new feature is expected to offer users a more efficient and intelligent way to create presentations, further accelerating the adoption of AI in professional settings.
10. Google Brings New AI Image Editing Tools to Slides and Vids
Google has announced two new Gemini-powered AI image editing features for its Google Slides and Google Vids platforms. The first feature allows users to completely replace an image’s background with a simple text prompt. The second feature allows users to extend an image’s background. Users select an image, click the “Generate image” icon in the side panel, and choose the extension function. They can then select a new aspect ratio, such as 16:9 or 4:3.
These new features are designed to help users manage image backgrounds more easily and efficiently.
Final Words on This Week’s AI News
The latest AI news shows a clear focus on making technology more personal and practical.
We’re seeing a shift from massive models to smaller, more efficient ones that can run on your phone, like Google’s Gemma 3. Companies are also creating more specialized tools to help with learning, creativity, and daily tasks, as seen with Anthropic’s “Learning Mode” and Google’s new editing features for presentations.
Overall, AI is becoming more integrated into our lives, and the industry is grappling with how to balance innovation with authenticity and a good user experience.
Read More:
- AI News from August 2–8, 2025
- AI News from July 26–August 1, 2025
- AI News from July 19–25, 2025
- AI News from July 12–18, 2025
