Anthropic has taken decisive action to prevent its AI models from being used for malicious purposes, including spying on ethnic minorities and dissidents in China, Iran, and West Africa. Between January and July, the company blocked operations targeting pro-democracy figures in Hong Kong and Iranian opponents abroad. Additionally, Anthropic stopped attempts to use its technology for biological weapon research and cyberattacks, warning that state actors are increasingly leveraging artificial intelligence for surveillance and harmful activities.
The company specifically reported that bad actors tried to use its chatbot, Claude, to research biological weapons over the past eight months. In one instance, a scientist requested assistance writing a grant to make the chikungunya virus more harmful to animals. Anthropic blocked these requests due to the potential for military misuse and plans to limit access to powerful models for trusted researchers only. The firm added stronger safeguards to its latest models to prevent such activities.
While Anthropic addresses safety, the broader AI market sees significant growth and regulatory scrutiny. Server sales rose by 15.4 percent in the second quarter, driven by strong demand for AI hardware. GPU-accelerated servers now account for nearly 53 percent of total revenue, with buyers expanding beyond large cloud providers to include enterprises and government programs. Major brands like Dell Technologies are gaining market share from smaller white box manufacturers.
Regulatory and ethical concerns continue to mount. Missouri Senator Josh Hawley has added OpenAI to his investigation into AI safety, demanding documents by September 8 following protests over license plate reading systems. Meanwhile, OpenAI CEO Sam Altman has questioned the legality of slowing down AI development, arguing it would be a huge problem for the world. Experts debate whether such slowdowns are necessary for safety or harmful to progress.
Other sectors are also integrating AI into daily operations. The Transportation Security Administration is adopting facial recognition and AI to speed up airport screenings, with Orlando International Airport already testing the technology. In healthcare, the 2026 China International Fair for Trade in Services showcased AI-powered tools like an embodied massage robot and the SupRing wearable device. FuriosaAI is expanding its presence in Singapore to sell RNGD inference accelerators, while SS&C Technologies launched Black Diamond AI to assist financial advisors with natural language data queries.
Key Takeaways
- Anthropic blocked government attempts to use its AI for spying on ethnic minorities and dissidents in China, Iran, and West Africa.
- The company stopped requests to use Claude for biological weapon research, including efforts to make the chikungunya virus more harmful.
- Server shipments increased by 15.4 percent in the second quarter due to surging demand for AI hardware.
- GPU-accelerated servers now represent nearly 53 percent of total server revenue.
- Missouri Senator Josh Hawley has added OpenAI to his investigation into AI safety and data storage practices.
- OpenAI CEO Sam Altman expressed concern that slowing AI development could create significant legal and global problems.
- The Pentagon is developing the Space Threat Intelligence Synthesis Engine to detect missile and space threats within seconds.
- The TSA is implementing facial recognition and AI to streamline airport security screenings.
- FuriosaAI opened a new Singapore hub to boost sales of RNGD inference accelerators in the Asia-Pacific region.
- SS&C Technologies launched Black Diamond AI to help financial advisors perform natural language data queries.
Anthropic blocks AI spying and weapon research
Anthropic announced it stopped several cases where governments used its AI models to spy on ethnic minorities and dissidents. The misuse came from China, Iran, and West Africa between January and July. These operations targeted groups like pro-democracy figures in Hong Kong and Iranian opponents abroad. The company also blocked attempts to use AI for biological weapon research and cyberattacks. Anthropic warned that state actors are increasingly using artificial intelligence for surveillance and harmful activities.
Anthropic warns of AI use for biological weapons
Anthropic reported that bad actors tried to use its AI chatbot Claude to research biological weapons over the past eight months. In one case, a scientist asked for help writing a grant to make the chikungunya virus more harmful to animals. The company blocked these requests because such research could be used for military purposes. Anthropic stated it cannot know if all such research is malicious, so it acts with caution. The firm plans to limit access to powerful AI models to trusted researchers only.
Anthropic stops bad actors from using Claude
Anthropic released a report showing it blocked foreign hackers from using its Claude AI to build cyberattacks and biological weapons. The company stopped attempts by groups linked to China and Iran to misuse its technology. White House tech advisor David Sacks noted that international rules to stop AI development are hard to enforce. Anthropic says it will continue to monitor and stop malicious use of its models. The report highlights how AI can be used for both good and dangerous purposes.
Anthropic blocks misuse of AI for weapons
Anthropic said it blocked bad actors from using its AI models for cyberattacks, surveillance, and biological weapon research. The company found requests to make the chikungunya virus more dangerous between December 2025 and August 2026. Anthropic added stronger safeguards to its latest models to prevent such harmful activities. The report includes examples of malicious code and prompts that the company stopped. The firm plans an initial public offering this fall while addressing safety concerns.
TSA uses AI and facial recognition for security
The Transportation Security Administration is adopting new technology to make airport security faster and safer. TSA Administrator David P. Cummins says the agency wants to use artificial intelligence and facial recognition for future screenings. Travelers might soon move through parts of security without showing ID or boarding passes. Some passengers worry about privacy and the need for human oversight in these systems. Orlando International Airport already uses facial recognition for certain international flights.
FuriosaAI opens Singapore hub for AI chip sales
FuriosaAI has launched a new subsidiary in Singapore to boost its sales of RNGD inference accelerators across the Asia-Pacific region. These chips provide high-performance computing for AI and machine learning while using very little power. The new office will offer sales and support services to data center operators and cloud providers. This move helps the company grow as demand for efficient AI solutions continues to rise.
David Gardner warns AI could make us lazy
David Gardner argues that the biggest danger of artificial general intelligence is not what it does to us, but what we do to ourselves. He believes AI might cause people to stop thinking critically because it provides instant answers for free. If knowledge becomes too easy to access, students may avoid the hard work of higher education and training. Gardner fears that without rigorous learning, humanity could lose the skills needed to understand how technology works.
AI healthcare robots and gadgets shine at CIFTIS
The 2026 China International Fair for Trade in Services in Beijing showcased many new AI healthcare products. Visitors saw an embodied intelligent massage robot that can copy human massage techniques almost perfectly. Another highlight was the SupRing, a wearable device from Beijing Tsinghua Changgung Hospital that tracks health data like heart rate and blood oxygen. An AI-powered fundus camera from Beijing Tongren Hospital and Airdoc also drew attention for detecting eye diseases and liver risks. These tools help doctors provide better care and reach more people.
Server sales grow despite high prices due to AI
Server sales are rising even though prices have gone up because of strong demand from AI projects. Shipments increased by 15.4 percent in the second quarter as companies and governments buy more hardware. GPU-accelerated servers, which are essential for AI, now make up nearly 53 percent of total revenue. Buyers are expanding beyond large cloud providers to include enterprises and government programs. Major brands like Dell Technologies are gaining market share from smaller white box manufacturers.
Pentagon uses AI to track space and missile threats
The Pentagon is developing new AI software to help identify space and missile threats faster than humans can. The project, called the Space Threat Intelligence Synthesis Engine, combines data from satellites, radar, and video feeds. This system aims to provide alerts within two to five seconds to help operators make quick decisions. It will work alongside human analysts and can even trigger automated defense responses. The goal is to improve speed and accuracy in detecting dangerous events during conflicts.
Missouri Senator Josh Hawley probes OpenAI over AI risks
Missouri Senator Josh Hawley has added OpenAI to his investigation into artificial intelligence safety. As the chair of the Senate Judiciary subcommittee on crime and counterterrorism, Hawley wants OpenAI to hand over documents by September 8. This action follows recent protests about camera systems that read and store license plate information. The senator is concerned about how these technologies are being used and stored by public employees.
SS&C Technologies launches Black Diamond AI for wealth management
SS&C Technologies Holdings has released Black Diamond AI to upgrade its Black Diamond Wealth Solutions platform. This new tool uses in-platform assistance and natural language data queries to help financial advisors. The system is governed by the SS&C AI Gateway and connects to the WorkHQ orchestration system. These features aim to reduce manual tasks and improve efficiency for large wealth management practices.
OpenAI CEO Sam Altman questions legality of AI slowdown
OpenAI CEO Sam Altman is concerned about the potential legal issues if the AI industry slows down. He stated in a recent interview that slowing AI development would be a huge problem for the world. The company faces challenges like a lack of diversity in research and a need for more funding. Experts debate whether a slowdown is necessary for safety or harmful to progress, but the legality remains unclear.
Sources
- China, Iran among countries that have used AI to aid spying, Anthropic says
- Anthropic warns of bids to use AI to build biological weapons
- Anthropic blocks bad actors from using Claude AI to build bioweapons
- Anthropic says it blocked misuse of its AI that could have supported biological weapons
- TSA looks to AI, facial recognition as airport security enters its next chapter
- FuriosaAI Launches Singapore Hub for Asia-Pacific AI Chip Sales
- David Gardner: The AI danger no one is talking about
- AI-powered healthcare products draw attention at 2026 CIFTIS
- Higher prices can't crimp server sales as AI drives demand
- Pentagon looks to AI to identify space and missile threats
- Josh Hawley adds OpenAI to his probe of artificial intelligence
- Did Black Diamond AI Launch Just Shift SS&C Technologies Holdings' (SSNC) Investment Narrative?
- OpenAI Wants to Know if an AI Industry Slowdown Would Even Be Legal
Comments
Please log in to post a comment.