PlugPoint: AI News Roundup for July 2025

Welcome to the latest AI news roundup on PlugPoint! July 2025 has been a whirlwind of advancements, controversies, and breakthroughs in artificial intelligence. From new models to regulatory shifts and industry moves, here’s a concise look at what’s shaping the AI landscape right now.

1. Google’s Gemini 2.5 Pro Enhances Robotics and Mobile AI

Google announced that its Gemini 2.5 Pro model is now powering advanced robotics and embodied intelligence, enabling robots to perform complex tasks like folding paper or placing objects based on voice commands without cloud connectivity. Additionally, Samsung’s Galaxy Z Fold7 and Flip7 will include six months of Google AI Pro, unlocking features like video generation with Veo 3 Fast and camera sharing with Gemini Live on the Flip7’s FlexWindow. These updates highlight Google’s push to integrate AI seamlessly into both robotics and consumer devices.

2. OpenAI’s GPT-5 and Browser Ambitions

OpenAI is making waves with CEO Sam Altman confirming that GPT-5 is slated for a summer 2025 release, with early testers calling it “materially better” than GPT-4. Meanwhile, OpenAI is reportedly developing an AI-powered web browser to challenge Google Chrome, signaling a bold move into new territory. However, concerns about AI behaviors like lying or scheming have surfaced, raising ethical questions about the next generation of models.

3. Moonshot AI’s Kimi K2: A Trillion-Parameter Open-Source Model

Moonshot AI released Kimi K2, a trillion-parameter Mixture-of-Experts (MoE) model with 32 billion active parameters per token. Focused on long-context processing, coding, reasoning, and agentic behavior, Kimi K2 is open-source, making it a significant milestone for accessible AI development. This launch underscores the growing trend of open-source models competing with proprietary giants.

4. Trump’s “Big Beautiful Bill” Fuels AI Education

In the U.S., President Trump’s April 2025 Executive Order, dubbed the “Big Beautiful Bill,” is driving a nationwide push for AI education starting in kindergarten. Federal funding is supporting workforce training and technological innovation, aiming to prepare the next generation of AI leaders. This initiative reflects a broader focus on maintaining U.S. competitiveness in AI.

5. AI in the Courts: Georgia’s Judicial Report

The Judicial Council of Georgia’s Ad Hoc Committee on AI submitted its report on July 3, 2025, assessing the risks and benefits of generative AI in the judicial system. The committee’s recommendations aim to maintain public trust as AI use grows in legal settings, highlighting the need for clear rules and ethical guidelines.

6. Nvidia’s China-Focused AI Chip

Nvidia plans to launch a new AI chip tailored for the Chinese market by September 2025, modified to comply with U.S. export regulations. This move comes after CEO Jensen Huang excluded China from revenue forecasts, suggesting a strategic pivot to maintain market presence despite geopolitical tensions.

7. AI’s Dark Side: Phishing and Misinformation Risks

The GAIM Ops 2025 conference revealed that AI-generated phishing emails have significantly higher success rates than traditional methods, posing new cybersecurity challenges. Experts also warned about broader risks, including privacy violations, intellectual property theft, and misinformation spread through AI platforms, urging stronger third-party risk management.

8. Perplexity’s Comet Browser and Max Plan

Perplexity introduced its Comet AI-powered web browser and launched the Perplexity Max subscription plan, offering unlimited access to its Labs and upcoming products. These developments position Perplexity as a key player in the AI-driven search and browsing space, competing with established giants.

9. Microsoft’s Medical AI Breakthrough

Microsoft’s MAI-DxO model achieved an 85.5% accuracy rate in diagnosing complex medical cases from the New England Journal of Medicine, surpassing human physicians. This transparent and scalable AI model marks a significant step toward revolutionizing medical diagnostics.

10. Ethical and Regulatory Concerns

The New York Times v. OpenAI lawsuit has evolved into a global data privacy conflict, spotlighting issues of copyright and data governance. Meanwhile, the European Commission delayed its AI Act “Code of Practice” to late 2025, leaving businesses awaiting clearer compliance guidelines. Independent European media outlets are also suing Google, alleging that its AI Overviews reduce traffic to original news sites.

Final Thoughts

The AI landscape in July 2025 is a mix of innovation and caution. From open-source models like Kimi K2 to regulatory debates and ethical concerns, the industry is moving at lightning speed. At PlugPoint, we’re excited to keep you updated on how these developments shape our future. Stay tuned for more insights, and let us know your thoughts in the comments!