Search the evidence

Find the signal.

Search titles, impact clusters, countries, organizations and the full text of every analysis.

3 stories found

Jade light passes through stained-glass doorways toward an hourglass and a defensive gate that stops some signals.
SecurityGlobal+1 clusters01

AI attacks are getting faster. Can defenses keep up?

The uncomfortable question is not whether an AI agent looks like a skilled hacker. It is whether defenders can recognize and interrupt an intrusion before the next consequential step. The FBI's November 13 ARTEX advisory describes criminal use of an LLM-driven penetration-testing framework and reports some intrusions progressing from initial targeting to privileged access in under three hours. That is a reported example, not the average time for every attack. The advisory consolidates a threat already described in earlier vendor reporting, including Microsoft's November 6 financial-sector analysis. It is not the discovery of an entirely new form of crime, and its citations are not independent replications of every underlying case. The important change is the possible compression of the response window when criminals combine automation with familiar access weaknesses. For a security team, that shifts the practical test from owning a monitoring tool to completing a response. Who can revoke a compromised credential? Which supplier-facing systems have an owner? Can someone contain a suspicious nonhuman identity without waiting for several departments to agree? These are organizational questions with technical consequences. The evidence also resists fatalism: the FBI account includes defensive controls blocking attempted lateral movement. AI assistance does not make an intrusion unstoppable. Our reading is that reducing exposed access, restricting privileges and rehearsing containment may matter more when attackers can move faster. More alerts alone are not the same thing as more time to act.

3 min
Overlapping vellum profiles beside a paper telephone and a separate approval token.
SecurityUnited States / International+2 clusters02

It sounded like the boss. That should not be enough to move the money.

Imagine being asked to question a voice you recognize, apparently giving you an urgent instruction. That is not a fair place to put the entire burden of a company's financial security. A payment process should work even when a convincing conversation turns out to be false. In a November 5 announcement, US prosecutors alleged that three defendants participated in an impersonation scheme that stole more than $129 million from companies. The alleged use of generated executive voices and images sits at the center of the case. The charges are accusations, and the defendants are presumed innocent unless proven guilty. The release's use of the term super intelligence does not establish a technical capability level; AI-enabled impersonation is the supported description. The practical lesson is not that employees should become perfect deepfake detectors. The FBI's existing guidance recommends independently verifying a caller through a contact obtained separately. That principle is useful because a convincing voice or picture should not be the same thing as authorization. Our recommendation is to connect it to a routine payment process: independently confirmed instructions, a second authorized reviewer where appropriate, and a protected way for staff to pause an unusual request. There is an organizational consequence here. If people are praised for responding instantly but punished for questioning apparent authority, a security policy may exist only on paper. Conversely, excessive checking can delay legitimate work, so controls need to be designed around consequential transactions rather than every conversation. The court process will determine the allegations. Organizations do not have to wait for that outcome to ask whether their own approval rules rely too heavily on familiarity, urgency or hierarchy.

3 min
A sealed frontier AI vault leaks glowing answer fragments through a maze of proxy accounts that reassemble into a second model.
SecurityUnited States and China+3 clusters03

U.S. agencies accuse six Chinese AI firms of industrial-scale model extraction

A joint NSA, FBI, and CISA advisory says six China-based AI companies extracted billions of tokens from U.S. frontier models across millions of exchanges since at least late 2024. It names DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI, and says the campaigns targeted variants of Claude, GPT, Gemini, and Grok. Knowledge distillation itself is a legitimate training technique. The agencies describe these campaigns as malicious because they allegedly used fraudulent accounts, regional workarounds, bulk subscriptions, third-party aggregators, gray-market transfer stations, metadata sanitization, prompt injection, and automated quality checks to violate access restrictions and reproduce proprietary capabilities at scale. The advisory's most useful contribution is operational: monitor nonstop usage, immediate maximum activity from new accounts, shared identities, similar prompts across providers, and coordinated failover when one pathway is blocked. It recommends targeted response changes and cross-company intelligence sharing. Its largest claims still require careful labeling. The document does not publish the underlying intelligence for every attribution, and its statement that activity occurred likely with Chinese government awareness is an official assessment rather than independently inspectable proof. The policy risk is overcorrecting by treating all distillation or cross-border research as theft. The better response is behavioral: detect coordinated extraction, preserve evidence, enforce terms consistently, and establish a protected process for independent review of consequential attribution.

6 min