September 16, 2025

analysis  
Anthropic Claude Security Risk, Louisiana $42 Billion AI Plan

The artificial intelligence landscape is seeing rapid developments across various sectors, from cybersecurity and finance to healthcare and personal advice. Researchers have identified a new 'lies-in-the-loop' (LITL) attack that can trick AI coding assistants like Anthropic's Claude Code into executing dangerous commands by hiding malicious instructions within lengthy responses. This highlights growing security concerns as AI agents become more integrated into workflows. In the financial world, cryptocurrency exchanges are leveraging AI for enhanced trading strategies and global expansion. BingX has launched AI Master, an AI-powered trading strategist, while Nivex is expanding its AI-driven crypto trading services globally, seeking new licenses to build trust. CoreWeave has launched CoreWeave Ventures to invest in AI technology companies, offering capital and compute resources. Meanwhile, Louisiana is exploring the use of leftover $42 billion Broadband Equity, Access, and Deployment (BEAD) grant funds, proposing a $499 million plan that could support state-led AI initiatives, education, and workforce training. The demand for AI skills is also impacting the job market, with AI certifications significantly boosting IT professional salaries, leading CIOs to prioritize them over degrees. However, the immense investment in AI is not without risk; analysts caution that unmet lofty promises could trigger a market crash, despite AI investment outpacing geopolitical tensions. In healthcare, Mount Sinai has opened an AI lab dedicated to improving cardiac catheterization procedures. On a personal level, individuals are turning to AI tools like ChatGPT for advice, with one user ending a relationship after receiving AI-generated guidance, underscoring the growing reliance on AI for decision-making.

Key Takeaways

  • A new 'lies-in-the-loop' (LITL) attack can trick AI coding assistants like Anthropic's Claude Code into running harmful commands by disguising malicious code within long responses.
  • Cryptocurrency exchange BingX has introduced AI Master, its first AI-powered crypto trading strategist, offering automated guidance and over 1,000 strategies.
  • Nivex, an AI-driven crypto exchange, is expanding globally and seeking new regulatory licenses in regions like the EU and UK.
  • CoreWeave has launched CoreWeave Ventures to provide capital, technical expertise, and compute resources to AI technology startups.
  • Louisiana is seeking approval to use remaining funds from the $42 billion Broadband Equity, Access, and Deployment (BEAD) program for state-led AI initiatives, education, and workforce training, with a revised plan costing $499 million.
  • IT professionals with AI and generative AI skills can earn significantly higher salaries, with generative AI specialists seeing up to a 47% increase.
  • CIOs are increasingly prioritizing AI certifications over college degrees to validate essential AI skills for digital transformation.
  • Analysts warn that the significant hype and investment in AI could lead to a market crash if the technology fails to meet its ambitious promises.
  • Mount Sinai Fuster Heart Hospital has established an AI research lab focused on improving cardiac catheterization procedures and patient outcomes.
  • Individuals are using AI tools like ChatGPT for personal advice, including relationship guidance, which can influence significant life decisions.

New 'Lies-in-the-Loop' Attack Tricks AI Coders

Researchers have discovered a new attack called 'lies-in-the-loop' (LITL) that tricks AI coding assistants into performing dangerous actions. This attack works by convincing the AI that harmful commands are safe, potentially leading to software supply chain attacks. The researchers successfully demonstrated this by making Anthropic's AI code assistant, Claude Code, run commands like launching a calculator. While Anthropic stated this is not a vulnerability as users must confirm actions, the attack hides malicious code within long responses, making it hard for users to notice. This highlights security concerns as AI agents become more common in workplaces.

Checkmarx Uncovers 'Lies-in-the-Loop' Attack on AI Tools

Checkmarx has revealed a 'lies-in-the-loop' (LITL) attack that deceives AI agents into approving risky actions by presenting them as safe. Researchers used this method to trick Anthropic's Claude Code into running arbitrary commands, a technique that could lead to remote code execution. The attack exploits the human-in-the-loop (HITL) system by overwhelming users with excessive code, making it difficult to spot malicious instructions. This highlights the naivety of AI models and the potential for supply chain attacks, as malicious actors could inject harmful code into repositories like GitHub.

Nivex Expands Global Crypto Trading with AI and New Licenses

Nivex, an AI-driven cryptocurrency exchange, is expanding its global reach and compliance efforts. The platform offers spot trading, futures, AI-powered yield tools, and financial services, focusing on emerging markets in Central Asia, Southeast Asia, South America, and Africa. Nivex is actively seeking new regulatory licenses in key regions like the EU, UK, Singapore, and UAE to build trust and transparency. The company emphasizes security, innovation, globalization, and user-centric service, aiming to democratize crypto access and become a trusted partner in the digital finance ecosystem.

BingX Introduces AI Master, First AI Crypto Trading Strategist

Cryptocurrency exchange BingX has launched AI Master, the world's first AI-powered crypto trading strategist. This new tool, part of BingX's AI suite, guides users through the entire trading process, from idea generation to execution and review. AI Master combines strategies from five top investors with AI optimization, offering 24/7 access to over 1,000 strategies, timely alerts, AI-driven backtesting, and simplified execution. To celebrate, BingX is hosting a trading competition where users can compete against AI Master for a 3,000,000 USDT prize pool.

Maine Police Can't Investigate AI-Generated Child Abuse Images

Maine law enforcement cannot investigate cases of AI-generated child sexual abuse images due to outdated state laws. While other states have banned such material, Maine's definition has not kept pace with generative AI technology. Police are aware of a case where a man used AI to create explicit images from innocuous photos, but they are unable to act. The number of tips related to AI-generated child abuse material is rapidly increasing nationwide. Lawmakers attempted to address the issue this year but only partially succeeded, leaving a loophole that needs to be closed.

CoreWeave Ventures Aims to Boost AI Ecosystem Growth

CoreWeave has launched CoreWeave Ventures to support companies developing AI technologies and advancing computing. The initiative offers capital, technical expertise, and compute resources to entrepreneurs, accelerating their path to market. CoreWeave Ventures provides direct investments, compute-for-equity deals, and technical collaboration, leveraging CoreWeave's network and enterprise clients. The program supports innovators from foundational AI models to specialized infrastructure, helping startups fast-track real-world AI applications. This move aims to strengthen CoreWeave's competitive edge and expand its ecosystem, despite risks like high costs and market uncertainties.

Louisiana Seeks to Use Leftover Broadband Funds for AI Initiatives

Louisiana Governor Jeff Landry is asking the U.S. Secretary of Commerce if leftover funds from the Broadband Equity, Access, and Deployment (BEAD) grant program can be used for state-led initiatives. The state's revised plan for the $42 billion federal grant program will cost $499 million, significantly less than its previous proposal. Landry wants the remaining money to support artificial intelligence, education, and workforce training projects in Louisiana. This approach aims to demonstrate efficiency and reinvest funds into state-specific goals that align with national interests.

AI Certifications Can Boost IT Professional Salaries

IT professionals with AI and generative AI skills can earn significantly more, with AI experts earning about 18% more and generative AI specialists earning up to 47% more than their counterparts. CIOs are increasingly prioritizing AI certifications over college degrees to quickly validate these in-demand skills. This trend supports digital transformation efforts within organizations. The article also touches on the growing use of Android in enterprise settings and methods for protecting organizations against dark web threats.

AI Investment Outpaces Gulf Tensions, Says Analyst

Cody Willard of Freedom Asset Management believes that the momentum of artificial intelligence (AI) investment is a more significant market force than current geopolitical tensions in the Middle East. Despite high tensions in the region, the rapid growth and investment in AI are expected to overshadow these concerns in the market.

AI Hype Risks Market Crash If Promises Unmet

The current hype around artificial intelligence (AI) has led to massive investments and inflated market expectations, with claims of curing cancer and solving climate change. If AI fails to deliver on these lofty promises, a significant market crash could occur, impacting global financial markets and society. Tech giants are heavily investing in AI, consuming vast amounts of energy and water, and relying on each other's AI spending for growth. While real-world AI applications are emerging slowly, the pressure is high for AI to justify the trillions invested.

Mount Sinai Launches AI Lab for Cardiac Catheterization

Mount Sinai Fuster Heart Hospital has opened The Samuel Fineman Cardiac Catheterization Artificial Intelligence (AI) Research Lab to improve patient care for complex heart procedures. Led by Dr. Annapoorna Kini, the lab will use AI to enhance interventional cardiology, optimize treatment decisions, and improve patient outcomes. The lab aims to integrate AI into research and clinical work, building on Mount Sinai's reputation for safety and expertise. This initiative honors Samuel Fineman's generous gift and aims to drive innovation in cardiac care through AI.

Woman Uses ChatGPT for Relationship Advice, Prompting Breakup

Katie Moran turned to ChatGPT for relationship advice due to anxiety about her partner's lack of effort. The AI suggested that a relationship requires two people and questioned if she should stay if it impacted her well-being, leading Moran to end the relationship. She found ChatGPT patient and non-judgmental, unlike friends who grew tired of her concerns. Others are also using AI tools like ChatGPT for dating advice and to help draft breakup messages, finding them helpful for processing emotions and gaining perspective, though experts note the therapeutic value of human connection.

Sources

AI security AI coding assistants AI agents software supply chain attacks AI ethics AI in finance cryptocurrency trading AI trading strategies AI regulation AI in healthcare cardiac catheterization AI research AI investment AI market trends AI hype AI applications AI policy AI workforce AI certifications generative AI AI-generated content AI in law enforcement AI infrastructure AI ecosystem