Microsoft Revokes Claude Code Licenses and Shifts Back to Copilot CLI

Microsoft has reportedly revoked, across the board, licenses for Claude Code, the Anthropic-built coding agent, that it had granted to thousands of its internal developers. The company is pushing migration to its own GitHub Copilot CLI, making its policy of reducing reliance on external tools unmistakably clear.

Some developers have voiced concerns that their tool choices are narrowing, exposing a gap between engineers on the ground who want the best tools for productivity and platform vendors pressing ahead with ecosystem lock-in.

arXiv Further Tightens Rules on AI-Generated Papers

The preprint server arXiv has made its rules on AI-generated content even stricter. The platform plans to impose stiffer penalties on AI usage that deviates from the core purpose of maintaining research quality, against a backdrop of unrelenting cases of unchecked AI-generated text slipping into submissions.

The declining quality of AI research papers has also been covered by The Verge, which points to the ironic situation that “AI papers are getting better, and that's becoming a big problem for scientists.” It underscores a structural challenge: the peer review system simply cannot keep up with AI-generated content.

Video AI Firm Runway Makes Full Entry into Japan with Over ¥6 Billion Investment

On May 14, US-based Runway announced its full-fledged entry into the Japanese market. The company's CEO said that “Japan has the most sophisticated creative industries in the world,” and Runway plans to invest more than ¥6 billion. As commercial use of video generation AI takes off in earnest, all eyes are on the impact on content production workflows.

Cerebras Doubles Its IPO Price: AI Chip Demand “Is Not Speculation”

AI chip venture Cerebras has raised its IPO price to double the original figure. Its CEO emphasized that “demand for AI chips is not speculative,” signaling that demand for data center AI semiconductors is steadily growing on the back of real usage.

Valuations of AI-related stocks in the IPO market are climbing, and NVIDIA's dominance as the sole powerhouse could start to shift.

Google AI Overviews Deals a Direct Blow to Publishers' Health Traffic

A study reportedly shows that Google's AI search feature, AI Overviews, is dealing a heavy blow to media traffic, particularly in the healthcare vertical. Because users get AI-generated answers right on the search results page, clicks through to the original articles are declining, raising concerns about the impact on publishers' revenue models.

Meta May Tap Employees' Mouse Movements and Keystrokes for AI Training

It has come to light that Meta plans to record employees' mouse movements and keystrokes and use them as AI training data. Privacy concerns over employee surveillance are mounting, once again raising the question of where the line should be drawn on corporate data use.

Kioxia Posts Record Results: ¥1 Trillion in Revenue, Net Profit Up 2,990%

Kioxia Holdings recorded revenue of ¥1,000.2 billion for the January–March 2026 quarter, with net profit up 2,990%. President Hiroo Ota described the results as “record revenue and profit growth,” driven by a surge in demand for NAND flash memory for data storage amid the AI boom. It's clear that the benefits are extending beyond AI semiconductors to data storage infrastructure as well.