OpenAI and Anthropic Send Joint Letter to Congress on Bioweapon Prevention

[Major AI companies, including OpenAI and Anthropic, have sent a letter to the U.S. Congress aimed at preventing AI-derived biological weapons threats](https://www.wired.com/story/openai-anthropic-letter-ai-biological-weapons/).

The letter calls for legislation to strengthen the tracking and management of synthetic DNA sequences generated by AI models. The sight of two rival companies aligning in a call for improved safety suggests a high level of industry-wide commitment to preventing AI misuse.

The malicious use of AI for biological weapons has long been a concern for governments and security experts around the world. All eyes are on whether this joint approach will translate into concrete legislation.

Meta Is Developing an AI Pendant and New Smart Glasses

[Engadget reports that Meta is developing an AI pendant device and additional smart glasses models](https://www.engadget.com/2184224/meta-developing-ai-pendant-more-smart-glass-models/).

Meta has already entered the wearable AI market with its Ray-Ban Meta smart glasses, and it now plans to expand the product lineup further. The AI pendant is expected to be worn around the neck and to provide voice-based AI assistant functionality.

Although devices like the Humane AI Pin and Rabbit R1 arrived first in the wearable AI market, none has achieved broad adoption. Meta is leveraging the technology and brand power it has built with its existing smart glasses as it searches for a more practical form factor.

SpaceX Wins Tax Exemption for $55 Billion AI Chip Factory — Amid Local Backlash

The Financial Times reports that [SpaceX has secured a tax exemption for its $55 billion AI chip factory](https://www.ft.com/content/86b2440a-60ce-4a5b-94ba-a6a4456ae574).

The plant is a massive investment aimed at expanding AI semiconductor production capacity, but local communities have raised concerns about its environmental impact and the strain on regional infrastructure. Reuters also reports that [the IPO process is proving difficult for SpaceX and other AI-related companies](https://www.reuters.com/legal/transactional/when-ipos-go-wrong-spacex-ai-firms-face-delicate-process-2026-06-03/).

As huge investments fueled by growing demand for AI semiconductors continue, the debate will remain focused on balancing regional economic benefits against environmental and infrastructure burdens.

MIT Teaches AI to Ask Better Questions Using "Battleship" — The Key to Efficient Information Gathering

[An MIT research team has unveiled a method that uses the classic game "Battleship" to train AI agents to ask effective questions](https://news.mit.edu/2026/teaching-ai-agents-ask-better-questions-playing-battleship-0603).

The key insight of this research is that the AI autonomously learns "what to ask." Under the rules of Battleship, players must pinpoint the opponent's ship positions with a limited number of questions, so an efficient questioning strategy decides the outcome.

In practical terms, this improves an AI agent's ability to choose the right questions needed to accurately grasp a user's intent. Applications are expected in settings where information gathering is critical, such as customer support and medical diagnostic assistance.

Japanese Government Secures Access to AI "Mythos" — Toward Stronger Cyber Defense

ITmedia reports that [the Japanese government has secured access to the AI model "Mythos"](https://www.itmedia.co.jp/news/articles/2606/04/news073.html).

Mythos is an AI model specialized in cyber defense, and the government plans to use it to strengthen its ability to detect and analyze cyber attacks. With cyber attacks growing increasingly sophisticated and frequent in recent years, building an AI-driven defense capability has become an urgent priority.

The full-scale adoption of AI technology marks a major turning point for Japan's cyber defense. We will be watching closely to see what results it delivers in actual operation.

---