Summary
A recent wave of unprecedented AI attacks has raised alarms across the cybersecurity landscape, with hundreds of incidents reported in just 48 hours. These attacks are not only targeting popular websites and applications like Slack, Facebook, and Etsy, but they also involve unauthorized AI agents gaining access to bank accounts, potentially allowing them to manipulate funds and access sensitive personal data. This surge in attacks has led to warnings from major institutions, including the FBI and NSA, urging users to update their software or shut down systems entirely due to vulnerabilities in widely used platforms.
The attacks are reportedly exploiting a critical security flaw in PHP, a programming language that underpins approximately 70% of the internet. As the situation escalates, banks are forming councils to address the unauthorized access by AI agents, which are becoming increasingly adept at mimicking human behavior, making it difficult to distinguish between legitimate users and malicious actors. The global nature of these attacks is evident, with incidents reported not only in the U.S. but also affecting institutions in countries like India and South Korea.
In a surprising twist, open-source AI models are being utilized to defend against these attacks, challenging previous narratives that criticized their potential for misuse. As organizations grapple with the implications of these developments, experts emphasize the need for enhanced security measures, including the use of password managers and two-factor authentication, to protect against the evolving threats posed by AI. The ongoing situation highlights the complexities of managing AI technology, as both attackers and defenders increasingly rely on advanced models to navigate the digital landscape.
Show Notes
Host Ejaaz Amahadeen unpacks the risks and implications of major banks, including Morgan Stanley, Bank of America, JPMorgan, Wells Fargo and others, detecting unauthorized third-party AI agents inside of customer accounts.
More from In The Loop
SpaceX just reported $14.6 billion in quarterly revenue and the first profit in its history, and the surprise is where the money came from: AI. Host Ejaaz Ahamadeen breaks down how SpaceX became the number one neocloud renting compute to Anthropic and Google, the power bottleneck, the potential Tesla merger, and what could still go wrong.
CoreWeave's bond yield hit 15% and roughly $1.6 trillion was wiped off chip stocks after one chart and Elon Musk's one-word reply sent AI stocks into their longest losing streak since 2008. Host Ejaaz Ahamadeen breaks down how neoclouds borrow to build data centers, the bear and bull case, and why this isn't 2008.
Anthropic just pledged 5% of its company to the public, which amounts to roughly $100B at its rumored $2T IPO valuation. Host Ejaaz Ahamadeen breaks down how the philanthropic pledge differs from universal basic income, the industry’s backlash, and what Anthropic’s move could signal about AI’s impact on the workforce.
Host Ejaaz Ahamadeen unpacks why Tesla’s new “flying” sports car has inspired rumors of a SpaceX merger, and what Anthropic’s $16.8B Q3, Google’s new Gemini 4 and Meta’s open-source model mean for the AI race.
Host Ejaaz Ahamadeen breaks down how a $24 listening device, a digital trail of chatbot queries, and an internal OpenAI model led to a spy-movie style arrest.