Florida woman arrested over chatbot messages Anthropic's Claude flagged as threats
Carli Heller, 30, was arrested in Florida after messages to Anthropic's Claude chatbot allegedly threatening the Lee County Sheriff's Office were flagged and passed to authorities. A legal commentator questions whether the messages count as a punishable threat.
1 / 1
The story, neutrally told
Left · 1Carli Michelle Heller, 30, of Florida was arrested after Anthropic's Claude chatbot flagged a conversation in which Heller allegedly made violent threats toward the Lee County Sheriff's Office. The VergeLC “Anthropic’s Claude AI chatbot flagged a conversation with 30-year-old Carli Michelle Heller after she allegedly made violent threats toward the Lee County Sheriff’s Office.” Read at The Verge ↗ Right · 1According to a probable cause statement quoted by Straight Arrow News, Heller wrote in late September that Heller was "going to shoot up the sheriff's right the f— now", and a day later wrote that it was "100% LAST CHANCE" and that Heller "GOT A NEW F—— GUN TODAY". ReasonRC “saying she was "going to shoot up the sheriff's right the f— now," she wrote to the phone app, according to a probable cause statement”“"100% LAST CHANCE" and that she "GOT A NEW F—— GUN TODAY."” Read at Reason ↗ Mixed · 2Human reviewers at Anthropic viewed the messages after the safety features triggered. The Verge, citing Gulf Coast News, says they reported the messages to police; Straight Arrow News, as relayed by Reason, says the review team escalated the risk to the FBI, which notified the sheriff's office. The VergeLC “Human reviewers then viewed the messages and reported them to police, the report says.” Read at The Verge ↗ ReasonRC “The team believed there was a big enough risk to escalate the risk to the FBI, which then notified the sheriff's office.” Read at Reason ↗
Right · 1Authorities arrested Heller on Sept. 30, charging Heller with making a written or electronic threat of a mass shooting or act of terrorism, a felony under Florida law. Heller pleaded not guilty. ReasonRC “Authorities arrested Heller on Sept. 30, charging her with making a written or electronic threat of a mass shooting or act of terrorism, a felony under Florida law” Read at Reason ↗ Right · 1Heller admitted writing the messages but denied any intent to carry them out, the report says. ReasonRC “Heller admitted writing the messages but denied any intent to carry them out, the report says.” Read at Reason ↗ Right · 1Eugene Volokh, writing in Reason, said they doubt this is a punishable threat under Florida Statutes § 836.10 unless there is evidence Heller realised at the time that a human would read the message and see it as a threat, citing T.R.W. v. State (Fla. App. 2023). ReasonRC “I doubt that this is a punishable threat, at least under the Florida law, Fla. Stats. § 836.10, unless there is evidence that the woman subjectively realized at the time that the message would be communicated to a human being who would perceive it as a threat.” Read at Reason ↗
Right · 1Volokh noted that Counterman v. Colorado (2023) lets states punish reckless threats, so Florida courts might allow that, though a 2026 appeals ruling, C.R. v. Burke, still follows the purpose-or-knowledge requirement. Volokh considered it more likely that Heller assumed the input would go only to the software, and said prosecutors would have to prove Heller's mental state beyond a reasonable doubt. ReasonRC “Florida courts thus may allow punishment for reckless threats, without violating the First Amendment—though they could also stick with T.R.W. and require purpose or knowledge.”“a prosecution would have to prove the woman's mental state beyond a reasonable doubt” Read at Reason ↗
Every sentence links to the reporting it rests on. The pill in front of each says where its sources sit: Left, Centre or Right when one side supplies at least half of them, Mixed when they are evenly split. The number is how many outlets it cites.
Left1 outlet
- Framing
- The Verge gives a brief news item on the arrest, framed around an AI chatbot flagging a user's conversation and human reviewers passing it to police.
- Emphasis
- The role of Claude and Anthropic's human reviewers in the report to police.
- Leaves out or plays down
- Does not give the quoted messages, the charge, the not-guilty plea, or any legal questions about the case.
- For example
-
“Human reviewers then viewed the messages and reported them to police, the report says.” — The Verge
Centre0 outlets
No centre outlet in our sources has covered this story yet.
Right1 outlet
- Framing
- Reason (Volokh Conspiracy) reposts Straight Arrow News's account, then adds a legal analysis asking whether a chatbot prompt can be a punishable threat.
- Emphasis
- Free-speech and mental-state doctrine: Florida's statute, T.R.W., Counterman, and the difficulty of proving Heller knew a human would read the message.
- Leaves out or plays down
- Gives little on Anthropic's review practices beyond the escalation, and no response from Anthropic or prosecutors.
- For example
-
“Simply typing a prompt into a chatbot, expecting it to be viewed only by the software, wouldn't qualify.” — Reason
What every side reports
- Carli Heller was arrested over messages to Anthropic's Claude chatbot allegedly threatening the Lee County Sheriff's Office.
- Human reviewers at Anthropic looked at the flagged messages before authorities were alerted.
Where accounts differ
-
Who Anthropic's reviewers reported the messages to
- Left
- The Verge, citing Gulf Coast News, says reviewers reported them to police.
- Right
- Reason, relaying Straight Arrow News, says they were escalated to the FBI, which notified the sheriff's office.
Anthropic organisation
Its safety features flagged the chats and its human review team judged the risk big enough to escalate.
“Anthropic's safety features triggered and sent the chats to its human review team to analyze.” — Reason
Carli Michelle Heller person
Admitted writing the messages but denied any intent to carry them out, and pleaded not guilty.
“Heller admitted writing the messages but denied any intent to carry them out, the report says.” — Reason
Lee County Sheriff's Office organisation
Was the target of the alleged threats and was notified by the FBI, according to Straight Arrow News.
“escalate the risk to the FBI, which then notified the sheriff's office.” — Reason
Left1 article
-
Florida woman arrested for allegedly making threats in an AI chat
Neutral Short factual report relaying Gulf Coast News about Claude flagging alleged threats and human reviewers alerting police.
Centre0 articles
No coverage yet.
Right1 article
-
ReasonRC ·
Can a Chatbot Prompt Be a Punishable Threat?
Critical Reposts a news account, then argues that typing a prompt into a chatbot is unlikely to be a punishable threat under Florida law.
- 5 Oct 16:11 First The VergeLC Florida woman arrested for allegedly making threats in an AI chat
- 8 Oct 15:59 +71h 48m ReasonRC Can a Chatbot Prompt Be a Punishable Threat?
Times are when each article was published, or when we first saw it if the outlet gave no time.