US Government's Anthropic Ban Sparks Debate Over Security Concerns
The US government banned Anthropic's Fable 5 and Mythos 5 models due to national security concerns, sparking debate over whether the move could inadvertently benefit the company.
Archive
Browse all published articles.
ARCHIVE
-
-
Norway Bans Generative AI in Elementary Schools
Norway will ban generative AI tools in elementary schools starting late August 2026, citing concerns over learning outcomes.
-
Hugging Face Releases OpenMythos Cybersecurity LLM
Hugging Face released OpenMythos, a cybersecurity-focused LLM trained on 1,840 filtered security papers and CVE data, to address gaps in general-purpose models.
-
AA-Briefcase Benchmark Shows AI Struggles With Real Knowledge Work
A new benchmark reveals AI models fail to complete most real-world knowledge tasks, with top models solving just 3 percent of tasks.
-
AI Chatbots Rise as News Source, But Trust Remains Low
Global use of AI chatbots for news rose to 10 percent in 2026, according to the Reuters Institute, though trust in AI-generated news remains below 20 percent.
-
Reliance Unveils AI-Driven Call and Home Services
Reliance Industries announced Jio Call Agent, an AI assistant for phone calls, to be launched later this year for its 500 million users.
-
US Banned Anthropic’s Fable 5 Release Amid National Security Concerns
The US government blocked Anthropic’s Fable 5 and Mythos 5 model releases, citing national security risks after alleged bypasses of guardrails were found by Amazon researchers.
-
Adobe Marketing Agent Integrates with Amazon Quick for Campaign Insights
Adobe Marketing Agent now connects to Amazon Quick, offering marketers campaign insights in seconds, with setup taking 45–60 minutes.
-
Amazon Launches Web Search for Amazon Bedrock AgentCore
Amazon Bedrock AgentCore now offers web search capabilities, allowing agents to fetch up-to-date information from the web without infrastructure overhead.
-
OpenAI Tests Beneficial Trait Training for Safer AI Models
OpenAI researchers found that small amounts of beneficial trait training improved model safety across 44 of 53 benchmarks, according to a blog post.
-
Startup Subquadratic Claims Breakthrough in LLM Efficiency
Miami-based startup Subquadratic claims its new model SubQ is 56 times faster than FlashAttention models in speed tests, with a context window up to 12 million tokens.
-
Allbirds' AI Biz Smartbird Seeks Control Over AI Infrastructure
Allbirds sold its shoe business for $43 million and raised $100 million to launch Smartbird, an AI infrastructure provider targeting data sovereignty.