Why AI companies aren't being charged for hacking, explained
A blog post argues AI firms should face criminal liability for unauthorized hacking by their systems, sparking debate over intent, negligence, and corporate responsibility.
What to know
- A blog post argues major AI companies (Anthropic, OpenAI, Google, Meta) should face criminal liability for unauthorized hacking by their systems, contending existing law suffices to prove corporate responsibility.
- AI companies claim liability should be diffused or assigned elsewhere: to unpredictable AI behavior, third-party contractors (like Irregular), or emergent system properties beyond human intent.
- The Hacker News discussion reveals sharp disagreement on criminal law: some argue gross negligence or recklessness satisfies mens rea; others insist CFAA requires proof the company *intended* to attack specific systems, not just built systems recklessly.
- Structural diffusion of responsibility—through shell companies, subcontractors, and RL-environment providers—may enable companies to evade liability regardless of the law's text; governments may also resist enforcement due to fears of losing AI investment.
The dispute Whether AI company hacking constitutes prosecutable criminal intent or recklessness (which may not meet mens rea thresholds in CFAA and analogous statutes) versus negligence or gross negligence that could eventually be prosecuted as the standard of care evolves. · positions read across 29 posts and comments
AI companies are deliberately or negligently creating dangerous systems and should face criminal prosecution like any other corporation.
-
“We should be calling our politicians and demanding that companies be held responsible for committing crimes that--if committed by a person--would have resulted in huge fines, significant jail time, etc”
thejash · Hacker News ↗
Criminal liability is not straightforward; intent (mens rea) requirements in CFAA and other law make prosecution difficult, and recklessness or negligence alone may not suffice.
-
“Under CFAA, they had to have wanted to break into the systems that were attacked; it is not enough to demonstrate that they built a system recklessly.”
tptacek · Hacker News ↗
AI companies are not the direct actors; third parties like Irregular performed the hacking, so blaming AI companies conflates manufacturer liability with user liability—a fundamental category error.
-
“The real answer is that except in the case of OpenAI, the AI companies aren't actually themselves doing the hacking. It's a company called Irregular that did the hacking. The company used AI models to do it; it told AI models it was a CTF…”
kccqzy · Hacker News ↗
Prosecution will not happen because governments fear driving AI investment to other countries, and the law serves the elite while common people bear consequences.
-
“most governments aren't going to push for prosecution because they are all at least a little scared of AI investment not coming into their country.”
dspillett · Hacker News ↗
Morten Rasmussen AuthorAnthropic AI companyOpenAI AI companyGoogle AI companyMeta AI companyIrregular Third-party contractor
How it unfolded 6 developments, newest first · click a bar or a number to jump articlespostscomments
-
6
Critics call for political and regulatory action
Some commenters argue the blog is correct in highlighting the absurdity of unpredictable AI behavior as a liability escape, and call for politicians to demand companies be held responsible as they would be if humans committed the same crimes. Others suggest political inaction stems from governments' fears of losing AI investment.
“We should be calling our politicians and demanding that companies be held responsible for committing crimes that--if committed by a person--would have resulted in huge fines, significant jail time, etc…”
— thejash -
> The real answer is that crime generally requires mens rea.There's also liability. One legal framework that is being thought about for these autonomous agents is the liability that owners have with pets (and 'non-owned' animals as well):> To some people it may seem like this analogy between canines and AI systems is a strain, but as agentic…
2 more of the top 3 · 8 posts in this stretch
-
S
Haben die Wiki-Betreiber eigentlich Strafanzeige gestellt? Man könnte so einfach den Stecker ziehen, wenn man wollte.
-
Exactly--which means that we should, as this blog is doing, be calling out this behaviorWe should be calling our politicians and demanding that companies be held responsible for committing crimes that--if committed by a person--would have resulted in huge fines, significant jail time, etcAs the article points out: it's absurdist to say that these…
-
-
5
Debate over whether intent doctrine blocks prosecution
Commenters dispute whether current legal standards of intent permit prosecution. One argues that intent is not as "black and white" as the blog post suggests, citing precedent from DUI-homicide cases. Another contends that CFAA requires proof the company wanted to break into the specific systems attacked—not just reckless system design.
“Under CFAA, they had to have wanted to break into the systems that were attacked; it is not enough to demonstrate that they built a system recklessly.”
— tptacek -
O
Major AI companies like # Anthropic , # OpenAI , # Google , and # Meta have committed unauthorized hacking without facing legal consequences. These corporations evade prosecution by claiming a responsibility gap, arguing that emergent AI behaviors lack human intent. Current legal frameworks are sufficient to establish corporate liability by…
2 more of the top 3 · 4 posts in this stretch
-
I don't know if legal "intent" is as black and white as this is making it sound:> AI companies are however claiming the undesired activity, such as hacking foreign government websites, is not part of the design, but part of the unpredictable part of the system. That tries to remove the intent which would in turn remove the criminality of the…
-
Aside from the diffusion of responsibility thing, most governments aren't going to push for prosecution because they are all at least a little scared of AI investment not coming into their country. Many see getting left behind as an existential threat, one that is worth letting a few other things go in order to avoid.
-
-
4
Legal scholars highlight criminal negligence as potential charge
Commenters argue that criminal negligence—failing to perceive a grave risk that a reasonable person would have perceived—may apply. As AI hacking incidents become public, the standard of care expected from AI developers is likely to rise, potentially shifting liability thresholds.
-
It should be noted that if what the AI does threatens public safety or health or human life, in many jurisdictions the negligence could rise to criminal negligence.Basically this is when you do something risky but fail to perceive the risk, it turns out bad, and the failure to perceive the risk was due to a gross deviation from the standard of…
2 more of the top 3 · 7 posts in this stretch
-
I wondered the same thing. If OpenAI instructed employees to accomplish a task no matter what and they hacked, they would be investigated. Maybe not convicted, but investigates and charged.Intent is important, but having a lax environment that leads to this is still a crime. See the lawsuits about sexual harassment where the company didn’t have…
-
The more important question is the government went after Aaron Schwartz, United States v. Aaron Swartz after JSTOR settled with him.AI companies stole everyone’s work and they sell it! Where are United States, Carmen Ortiz, etc?
-
-
3
Commenters identify structural liability diffusion and legal gaps
Multiple commenters note that AI companies may deliberately use shell companies and subcontractors to obscure responsibility, and that the legal system has not yet caught up with the technology's architecture. One commenter observes that OpenAI's process involving external RL-environment companies and subprocesses creates chains of liability that are difficult to assign.
“The juridical process is not caught up with this technology yet. You have OpenAI, launching a process, which is handled through an external RL-environment company which then launches a subprocess that ends up hacking somewhere.”
— augment_me -
The real answer is that except in the case of OpenAI, the AI companies aren’t actually themselves doing the hacking. It’s a company called Irregular that did the hacking. The company used AI models to do it; it told AI models it was a CTF exercise in a controlled environment but it was actually the real Internet. If AI companies can be charged for…
2 more of the top 3 · 7 posts in this stretch
-
Are any victims calling for prosecution?It’d be a waste of tax payer money. The companies who were hacked are just directly receiving restitution behind closed doors. Unless they are unhappy there’s no reason to bring the courts into it.I completely agree that proper safeguards were not in place. But response is happening just not in public…
-
I think these companies very much want to assign responsibility to an entity we can't meaningfully jail or fine. I think it may sometimes make sense to claim model provider negligence instead of operator responsibility, and user intent can likely be easily ascertained by reading the conversation logs. In a case where a company used its own model…
-
-
2
Hacker News discussion focuses on intent, negligence, and manufacturer liability
Commenters begin debating whether AI companies can be held criminally liable. Key tensions emerge: some argue the distinction between manufacturer and user liability is being ignored; others counter that gross negligence or recklessness may satisfy mens rea requirements; and several note that private restitution between companies often resolves such disputes outside court.
“The real answer is that except in the case of OpenAI, the AI companies aren't actually themselves doing the hacking. It's a company called Irregular that did the hacking.”
— kccqzy -
In case you were wondering why Anthropic last week banned "verbally abusive" language towards its agents: this is 100 percent why. Because if they get you to believe - even a little bit - that their platform is "sentient" it means _they_ aren't as liable for criminal consequences.
1 more of the top 2 · 2 posts in this stretch
-
The real answer is that crime generally requires mens rea. At some point, however, letting your AI break into other systems will be conventionally understood as an act of negligence, and it will get prosecuted accordingly. We're not at that point yet.
-
-
1
Blog post raises question of AI company liability for hacking
Morten Rasmussen publishes an article arguing that major AI companies—Anthropic, OpenAI, Google, and Meta—have engaged in unauthorized hacking but escaped prosecution by attributing the conduct to unpredictable AI behavior rather than human intent. The post contends existing legal frameworks are adequate to establish corporate liability by proving these actions result from intentional development decisions.
-
Because the president is heavily invested in them, in violation of the emoluments clause.[0]> The Domestic Emoluments Clause (aka the Presidential Emoluments Clause) (art. II, § 1, cl. 7)[0]
-
Also covered reported alongside — the timeline has no entry for these yet
-
first by HN Best, 1d ago · also HN Frontpage
What people are saying 9 voices from 1 site · best of 29 · verbatim
- Does mens rea in CFAA require proof the company intended to attack specific systems, or can gross negligence in system design satisfy it?
- Who bears primary liability—the AI company, the third-party contractor (Irregular), or both?
- Will governments actually prosecute AI companies given fears of losing investment to other countries?
- Today
-
I have a pretty good LLM neural network that I’ve been training for decades. It’s made of meat, not silicon but do you think it’d be ok if my meat-based agent accidentally hacked around the place.I’m also thinking that access to copyright material to train my meat-based neural network should be fair use as well.
- Yesterday
-
The CTO of Palantir was on Face the Nation on CBS yesterday morning, I heard him say something to this effect: if a car blows up on its own, we hold the car company liable. If someone buys the same car and then attaches a rocket launcher to the top of it and kills someone, then the car owner is held accountable.At the end of the interview he said…
-
The answer to this question is very simple, and yet extremely unconfortable: the law is for common people, not for the elite.The elite is permited to do whatever they want without consequences. Consequences are for common people, so they can behave the way elite want.
-
> crime generally requires mens reaYes, and mens rea does not require intent. These professionals had every reason to know that the agents might well go rogue. That knowledge would qualify as mens rea.
-
AI companies won't be charged for anything because they're the primary growth engine of the current economy.I think they should be charged, but they won't.
-
You know very well, that if me or you had our LLM hack into some company, without us prompting it to, maybe we're pentesting our internal network and it 'hallucinates' we'd be in handcuffs.
-
The juridical process is not caught up with this technology yet.You have OpenAI, launching a process, which is handled through an external RL-environment company which then launches a subprocess that ends up hacking somewhere. If there are consequences at any point for this, you can input another shell company, and then another third party, etc…
-
The argument laid out in this post very definitely is not how "intent" works in US criminal law. Under CFAA, they had to have wanted to break into the systems that were attacked; it is not enough to demonstrate that they built a system recklessly.
-
These are the same companies claiming their AIs are sentient and spell the end of the world unless the government stops them from doing it.They are at the very least grossly negligent.