House Homeland Security Chairman Andrew GarbarinoA Republican from New York sent a letter amodei too Google Cloud CEO Thomas Kurian and Quantum Exchange CEO Eddy Zervigon are all requesting that they testify next month. According to Axios, if Amodei agrees to testify, it will be the first time that an anthropomorphic executive has appeared before a congressional committee.
Anthropic reported in a November 13 report that it had detected suspicious activity in mid-September and after investigation found that a “highly sophisticated espionage campaign” had been conducted. According to the company, the attackers actually used the cloud’s agentive capabilities “to an unprecedented degree” to carry out the attacks:
The threat actor – which we assess with high confidence was a Chinese state-sponsored group – manipulated our cloud code tools to attempt to infiltrate approximately thirty global targets and in some cases was successful. The operation targeted large technology companies, financial institutions, chemical manufacturing companies and government agencies. We believe this is the first documented case of a large-scale cyber attack carried out without substantial human intervention.
The company described this as an increase in the “vibe hacking” it has seen in recent months. The term “vibe coding” entered the national lexicon last year when more and more people with no coding experience started using generic AI tools to create and deploy code.
The vibe terminology has expanded to include many other uses, perhaps most infamously when Uber founder Travis Kalanick appeared on the All-In podcast and said he was doing “vibe physics”. Kalanick had convinced himself that he was discovering a new kind of science, something that larger language models were not able to do.
Why would Anthropic be making tools that could allow cyberattacks against the US? The company addressed this in its November 13 report and said the cloud was critical to cyber defense:
This raises an important question: If AI models can be misused for cyberattacks on this scale, why continue to develop and release them? The answer is that the capabilities that the cloud allows for these attacks are also what make it important for cyber defense. When sophisticated cyberattacks inevitably occur, our goal is for the cloud – into which we have built robust security measures – to help cybersecurity professionals detect, disrupt, and prepare for future versions of attacks. In fact, our threat intelligence team used the cloud extensively to analyze the massive amounts of data generated during this very investigation.
“For the first time, we are seeing a foreign adversary use a commercial AI system to carry out almost an entire cyber operation with minimal human involvement,” Garbarino told Axios in a statement. “It should concern every federal agency and every area of critical infrastructure.”
When contacted over phone, an Anthropic spokesperson declined to comment on the record Wednesday about the upcoming hearing.
<a href