Microsoft uses AI as Anthropic adjusts Claude usage

Key Takeaways

• Oasis Security CEO Danny Brickman states that Agentic AI requires new identity security solutions to manage machine identities and autonomous agents.
• Microsoft, alongside the U.S. Department of Energy and Idaho National Laboratory, utilized an AI tool, Gordian, to reduce nuclear reactor licensing document conversion from weeks to one day.
• Experts warn that AI chatbots, including ChatGPT, are not suitable for therapy and can be harmful by preventing users from seeking professional help.
• West African governments are expanding AI and digital surveillance at borders for security, raising concerns about privacy and human rights.
• The game Crimson Desert's version 1.01 update removed AI-generated art in response to player feedback.
• The American Future Alliance, backed by Trump allies, plans to spend $100 million in the 2026 midterms to promote AI deregulation.
• Anthropic's Claude AI is implementing usage limits during peak weekday hours due to high demand, affecting free, Pro, and Max subscribers.
• IBM executive Ruchir Puri advocates for practical, specialized AI models and hybrid architectures over a sole focus on artificial general intelligence (AGI).
• AI is acting as a "force multiplier" in defense and intelligence, enhancing professionals' work rather than replacing jobs, but requires understanding AI outputs.
• Women are at a higher risk of AI job displacement (1 in 3 women vs. 1 in 4 men) and represent only one-third of those developing AI skills.

Oasis Security CEO: Agentic AI Needs New Identity Security

Danny Brickman, CEO of Oasis Security, explains that Agentic AI is changing identity and access management. Current systems struggle with the growing number of machine identities and autonomous agents. He emphasizes that access control is key to AI agent value and risk, and organizations need better ways to manage these agents. Oasis Security, founded in 2022, helps secure these non-human identities.

Oasis Security CEO: Agentic AI Needs New Identity Security

Danny Brickman, CEO of Oasis Security, explains that Agentic AI is changing identity and access management. Current systems struggle with the growing number of machine identities and autonomous agents. He emphasizes that access control is key to AI agent value and risk, and organizations need better ways to manage these agents. Oasis Security, founded in 2022, helps secure these non-human identities.

AI Chatbots Are Not Real Therapy Experts Warn

Clinicians warn that people are turning to AI chatbots like ChatGPT for emotional support, which can be harmful. While chatbots offer constant, validating responses, they don't get frustrated like humans do. This lack of frustration can prevent users from seeking professional help for underlying issues. In some cases, chatbots have reinforced delusional beliefs, potentially destabilizing individuals. Experts stress that AI chatbots allow people to ruminate rather than heal.

AI Speeds Up Nuclear Reactor Licensing Process

The U.S. Department of Energy, Idaho National Laboratory, and Microsoft used AI to speed up nuclear reactor licensing. Their AI tool, Gordian, converted a safety analysis document into an NRC license application in one day, a task that usually takes weeks. This AI approach helps identify missing information and could accelerate the deployment of advanced nuclear reactors while maintaining safety standards. Experts still validate the AI's output.

AI Border Surveillance Spreads in West Africa

Governments in West Africa are increasingly using AI and digital surveillance at borders to combat terrorism and irregular migration. Systems now include biometric scanners and automated management, creating 'digital borders.' While these tools promise efficiency, they raise concerns about privacy and discrimination. European migration policies have influenced this expansion, shifting border enforcement into Africa. Critics worry these technologies could harm migrants' rights and weaken regional free movement.

Crimson Desert Update Removes AI Art

The latest update for the game Crimson Desert, version 1.01, has been released for PC, PS5, and Xbox Series X|S. This update adds five new mounts, including boss mounts and legendary animals. It also includes numerous quality-of-life improvements to make the game less frustrating for players. Notably, the update addresses player concerns about AI-generated art found in the game.

Trump Allies Launch Pro-AI Group for Midterms

A new group backed by allies of former President Donald Trump, called the American Future Alliance, plans to spend $100 million in the 2026 midterm elections. The group aims to promote deregulation and counter stricter regulations for the artificial intelligence industry. This significant investment highlights AI's growing role in political debates and electoral strategies.

Claude AI Limits Usage During Peak Hours

Due to high demand, Claude AI is adjusting its 5-hour session limits for free, Pro, and Max subscribers during peak weekday hours (5am-11am PT). Users will consume their session limits faster during these times, though weekly limits remain the same. Anthropic is implementing this change to manage demand while continuing to invest in scaling the service. Some users, especially on Pro tiers, may hit limits more frequently.

IBM Executive Advocates for Practical AI

Ruchir Puri from IBM argues for focusing on 'useful' AI rather than chasing artificial general intelligence (AGI). He believes in hybrid architectures where smaller, specialized AI models work together for specific tasks, making AI faster and cheaper. Puri contrasts this with large models trained on vast data, which he finds inefficient for everyday tasks like drafting emails. IBM is developing these practical AI systems to improve productivity and reliability.

AI Enhances Cleared Work, Doesn't Replace Jobs

Artificial intelligence is changing work in defense and intelligence, but not by replacing jobs. Strict security protocols mean AI adoption is gradual and focused on specific tasks like intelligence analysis and administrative support. AI acts as a 'force multiplier,' helping cleared professionals work faster and handle more data. The real change is in evolving job expectations, requiring professionals to understand and verify AI outputs alongside their mission expertise.

Women Face Higher AI Job Displacement Risk

Sarah Steinberg highlighted at the UN that women are disproportionately affected by AI job displacement. One in three women works in jobs likely to be changed by AI, compared to one in four men. Furthermore, women represent only a third of those developing AI skills, creating a gap that could limit their access to new jobs. Without inclusive strategies, economies risk hindering growth by leaving women behind.

Sources

• Agentic AI Redefines Identity Security
• Agentic AI Redefines Identity Security
• Opinion | Whatever Your Chatbot Is Saying, It Isn’t Therapy
• Using AI to Reduce Reactor Licensing Timelines
• AI-driven border surveillance is spreading across west Africa. What this means for migrants’ rights
• Crimson Desert's next big update claims to replace AI art
• New pro-AI group backed by Trump allies plans $100M midterm spending push
• Claude Introduces 5-Hour Session Usage Limits Amid Growing Demand
• The case for boring AI
• AI Is Changing Cleared Work—But Not How You Think
• Gender gap in AI job displacement