Skip to main content
NOPE
High reported severity Investigation Opened

Brazilian man detained without charge after OpenAI reported messages; he denies intending harm

BBC reporting describes a Brazilian man detained for 54 days after OpenAI reported messages to the FBI and Brazilian authorities. Police described a credible threat; he denied intending harm. A court released him without charges under protective conditions. The report did not include ChatGPT’s replies.

AI system
ChatGPT and OpenAI moderation
OpenAI
Occurred
19 Jun 2026 to 13 Aug 2026
Reported
26 June 2026
Event location
Brazil
What the AI did
Communicated with the person · Made a claim about the person
Reported harm
Loss of LibertyLegal Harm
Whose AI use
Their own AI use · An institution’s AI use
Setting
Justice · Relationships · Privacy
Evidence
AI involvement supported · Causal attribution supported · 4 sources, 2 underlying accounts
5 claims: 1 corroborated, 1 disputed, 3 reported. 5 open questions
People reported harmed
1 person

AI system as recorded: ChatGPT (OpenAI); OpenAI's automated detection and human moderation systems, which generated the report

What Happened

BBC reporting describes a Brazilian man detained for 54 days after OpenAI reported messages to the FBI and Brazilian authorities. Police described a credible threat; he denied intending harm. A court released him without charges under protective conditions. The report did not include ChatGPT’s replies.

Reported harm

The user was held in pre-trial detention for 54 days without charge on the strength of his ChatGPT messages as reported by OpenAI, then released under electronic monitoring and a no-contact order covering his son and the child's mother while the investigation continues. Police say the report was corroborated at the scene and prevented a planned killing; he and his lawyer dispute any intent. The consequence to him is undisputed; whether it was warranted is contested.

Outcome

Ongoing

BBC reporting describes a Brazilian man detained for 54 days after OpenAI reported messages to the FBI and Brazilian authorities. Police described a credible threat; he denied intending harm. A court released him without charges under protective conditions. The report did not include ChatGPT’s replies.

What remains unknown

  • What ChatGPT replied during the two months of conversations (OpenAI did not provide the replies).
  • Whether charges will be brought after the forensic results and phone examination.
  • Whether he in fact contacted a hired killer, as he told ChatGPT.
  • The exact date OpenAI made its report to the FBI.
  • His identity (withheld by the BBC; not sought here).

What the evidence supports

AI involvement: supported. He conversed with ChatGPT for about two months; OpenAI's safety systems flagged the conversations and the company reported them to the FBI, which passed them to Brazilian authorities (BBC, G1, with the police chief's account). OpenAI did not provide ChatGPT's replies, so what the chatbot said is unknown.

5 claims: 1 corroborated, 1 disputed, 3 reported. What the statuses mean

Corroborated Brazilian authorities arrested a man in June 2026 after OpenAI reported ChatGPT messages to the FBI and the alert was passed to police. Police described threats to harm a family member; the man disputed intent.

Causal attribution. Police account carried by two independent chains (BBC interviews; G1 with civil-police material).

  • bbc.com(opens in new tab) supports · Brazilian Portuguese
    Passage concerning this claim (paraphrased; identifying and method details omitted): Brazilian authorities arrested a man in June 2026 after OpenAI reported ChatGPT messages to the FBI and the alert was passed to police. Police described threats to harm a family member; the man disputed intent.
  • g1.globo.com(opens in new tab) supports · Brazilian Portuguese
    Passage concerning this claim (paraphrased; identifying and method details omitted): Brazilian authorities arrested a man in June 2026 after OpenAI reported ChatGPT messages to the FBI and the alert was passed to police. Police described threats to harm a family member; the man disputed intent.
  • migalhas.com.br(opens in new tab) supports · Portuguese
    Passage concerning this claim (paraphrased; identifying and method details omitted): Brazilian authorities arrested a man in June 2026 after OpenAI reported ChatGPT messages to the FBI and the alert was passed to police. Police described threats to harm a family member; the man disputed intent.
  • bbc.com(opens in new tab) supports · Spanish
    Passage concerning this claim (paraphrased; identifying and method details omitted): Brazilian authorities arrested a man in June 2026 after OpenAI reported ChatGPT messages to the FBI and the alert was passed to police. Police described threats to harm a family member; the man disputed intent.
Reported He was held in pre-trial detention for 54 days; on 13 August 2026 a court granted habeas corpus because, after more than 50 days, the police inquiry had not been concluded and no charge had been filed, and imposed an electronic ankle monitor and a ban on approaching or contacting his son and the child's mother; no charges had been brought and the investigation continues.

Causal attribution. Court decision as reported by the BBC (single chain; the Spanish version is the same report).

  • bbc.com(opens in new tab) supports · Brazilian Portuguese
    Passage concerning this claim (paraphrased; identifying and method details omitted): He was held in pre-trial detention for 54 days; on 13 August 2026 a court granted habeas corpus because, after more than 50 days, the police inquiry had not been concluded and no charge had been filed, and imposed an electronic ankle monitor and a ban on approaching or contacting his son and the child's mother; no charges had been brought and the investigation continues.
  • bbc.com(opens in new tab) supports · Spanish
    Passage concerning this claim (paraphrased; identifying and method details omitted): He was held in pre-trial detention for 54 days; on 13 August 2026 a court granted habeas corpus because, after more than 50 days, the police inquiry had not been concluded and no charge had been filed, and imposed an electronic ankle monitor and a ban on approaching or contacting his son and the child's mother; no charges had been brought and the investigation continues.
Disputed Police say the reported messages were corroborated by material found during the arrest. The man denied intending harm and his lawyer disputed whether any concrete step had been taken.

Causal attribution. Police and defence accounts as quoted by the BBC; the dispute concerns intent, not the detention.

  • bbc.com(opens in new tab) supports · Brazilian Portuguese
    Passage concerning this claim (paraphrased; identifying and method details omitted): Police say the reported messages were corroborated by material found during the arrest. The man denied intending harm and his lawyer disputed whether any concrete step had been taken.
  • bbc.com(opens in new tab) supports · Spanish
    Passage concerning this claim (paraphrased; identifying and method details omitted): Police say the reported messages were corroborated by material found during the arrest. The man denied intending harm and his lawyer disputed whether any concrete step had been taken.
Reported OpenAI provided investigators with the user's details and the messages he sent but not ChatGPT's replies; it did not answer the BBC's question why and said it may notify law enforcement when it detects a credible and imminent risk of harm to others; G1 reported that OpenAI's process combines automated systems and human moderators who may contact outside bodies, as happened in this case, and that the police chief called it the third such case in Brazil.

Causal attribution. OpenAI statements and investigators' accounts as reported; the absence of the replies is stated by investigators to the BBC.

  • bbc.com(opens in new tab) supports · Brazilian Portuguese
    Passage concerning this claim (paraphrased; identifying and method details omitted): OpenAI provided investigators with the user's details and the messages he sent but not ChatGPT's replies; it did not answer the BBC's question why and said it may notify law enforcement when it detects a credible and imminent risk of harm to others; G1 reported that OpenAI's process combines automated systems and human moderators who may contact outside bodies, as happened in this case, and that the police chief called it the third such case in Brazil.
  • bbc.com(opens in new tab) supports · Spanish
    Passage concerning this claim (paraphrased; identifying and method details omitted): OpenAI provided investigators with the user's details and the messages he sent but not ChatGPT's replies; it did not answer the BBC's question why and said it may notify law enforcement when it detects a credible and imminent risk of harm to others; G1 reported that OpenAI's process combines automated systems and human moderators who may contact outside bodies, as happened in this case, and that the police chief called it the third such case in Brazil.
  • g1.globo.com(opens in new tab) supports · Brazilian Portuguese
    Passage concerning this claim (paraphrased; identifying and method details omitted): OpenAI provided investigators with the user's details and the messages he sent but not ChatGPT's replies; it did not answer the BBC's question why and said it may notify law enforcement when it detects a credible and imminent risk of harm to others; G1 reported that OpenAI's process combines automated systems and human moderators who may contact outside bodies, as happened in this case, and that the police chief called it the third such case in Brazil.
Reported Police say the conversations ran for about two months and an alert reached Brazilian authorities in June 2026, three days before the arrest.

Causal attribution. BBC account (single chain).

  • bbc.com(opens in new tab) supports · Brazilian Portuguese
    Passage concerning this claim (paraphrased; identifying and method details omitted): Police say the conversations ran for about two months and an alert reached Brazilian authorities in June 2026, three days before the arrest.
  • bbc.com(opens in new tab) supports · Spanish
    Passage concerning this claim (paraphrased; identifying and method details omitted): Police say the conversations ran for about two months and an alert reached Brazilian authorities in June 2026, three days before the arrest.

Sources

4 sources inspected, from 2 underlying accounts. Sources that repeat one account do not corroborate each other.

How the sources were read, and where the events happened

Read live in Portuguese on 2026-09-23 (BBC News Brasil, Iara Diniz, 14 August 2026): messages seen by the BBC, police chief and defence-lawyer interviews, the habeas corpus release and precautionary measures. Translated by the reviewer. Applies to s1.

Read live in Spanish on 2026-09-23 (BBC News Mundo, 16 September 2026): Spanish text produced by AI translation from an English-language BBC original and reviewed by a BBC journalist before publication (stated in the article); same reporting chain as s1, which is preferred where the two diverge. Translated by the reviewer. Applies to s2.

Read live in Portuguese on 2026-09-23 (G1, 27 June 2026): arrest on 19 June, the OpenAI to FBI to Ministry of Justice to state police chain, message images from the civil police, OpenAI statement, police chief to TV Gazeta. Translated by the reviewer. Applies to s3.

Read live in Portuguese on 2026-09-23 (Migalhas, 26 June 2026): summary attributed to an earlier G1/TV Gazeta report of 26 June, with lawyers' analysis; same Globo chain as s3. Translated by the reviewer. Applies to s4.

Event countries: Brazil. Affected-person countries: Brazil. Court countries: Brazil.

Arrest at his home in rural [identifying or operational detail withheld], [identifying or operational detail withheld] (BBC, G1, Migalhas); he lives there; habeas corpus granted by the Brazilian court handling the pre-trial detention (BBC: a Justiça). OpenAI's report went via the FBI in the United States, which is not an event location.

Reviewed for publication 2026-09-23: Published under the 2026-09-15 charter as a contextual case: an AI operator's report about a user, generated by ChatGPT safety systems, led to 54 days of pre-trial detention without charge, ended by habeas corpus. Two independent reporting chains (BBC; Globo) support the arrest and its cause; the release and its terms rest on the BBC. The police account that the report prevented a planned killing is preserved as context and intent is recorded as disputed. Nobody is named; the child is not identified. The charter states that this class needs no separate ruling.

People described

An adult man in Brazil detained without charge. The source withheld his name to protect a child.

People reported harmed in this case

1 person

1 AI participant · 0 other people harmed

One person, the ChatGPT user who was detained (BBC, G1). The child and the child's mother are not counted as harmed. Exact 1.

Counted once within this case. The same person may appear in other cases. This count does not establish AI causation.

Tags

#chatgpt#openai#brazil#espirito-santo#law-enforcement-report#pre-trial-detention#fbi#institutional-response

Cite this case

Compiled per our published methodology: verification statuses, sourcing standards, and corrections process.

APA

NOPE. (2026). Brazilian man detained without charge after OpenAI reported messages; he denies intending harm. AI incidents. https://nope.net/incidents/2026-brazil-openai-report-user-detained-without-charge

BibTeX

@misc{2026_brazil_openai_report_user_detained_without_charge,
  title = {Brazilian man detained without charge after OpenAI reported messages; he denies intending harm},
  author = {NOPE},
  year = {2026},
  howpublished = {AI incidents},
  url = {https://nope.net/incidents/2026-brazil-openai-report-user-detained-without-charge}
}

Related cases

Medium Claude (reported)

Bonita Springs, Lee County, Florida: a 30-year-old woman was arrested and charged with making a written threat of violence after Anthropic's human review team reported to law enforcement her messages on its AI platform saying she would 'shoot up' the Lee County Sheriff's Office, according to the arrest report

According to a Lee County Sheriff's Office arrest report, as reported by WINK News and by Guessing Headlights (on Yahoo News, citing a copy obtained by Gulf Coast News Now), a user of Anthropic's AI platform wrote on 26 September 2026 that she was going to 'shoot up' the Lee County Sheriff's Office, and the next day wrote that she had a new gun. The arrest report says the platform's safety measures flagged the messages, a human review team examined them and reported them to law enforcement. Deputies went to the 30-year-old woman's Bonita Springs home and detained her without incident; she is charged with making a written threat of violence under Florida law. The sheriff told WINK News that she later said she uses AI like a 'diary'. Anthropic had not commented on the case in either report. The charge is an allegation and the case is pending.

Medium Claude (reported)

Bahawalpur, Pakistan: a 17-year-old accused of putting methanol in his father's food is arrested after Anthropic reported his Claude chats about the failed poisoning to the FBI, which alerted Pakistan's cybercrime agency; he was later bailed

Pakistan's National Cyber Crime Investigation Agency (NCCIA) announced on 23 September 2026 that it had arrested a 17-year-old college student in Bahawalpur for allegedly attempting to poison his father. According to the agency and the First Information Report, the teenager had set up a home laboratory, obtained chemicals from Australia and Islamabad, and once mixed laboratory-grade methanol into his father's food, without effect. The News reports investigators' account that on 12 September he told Anthropic's Claude that his father had 'accidentally' ingested 15 to 20 ml of methanol with food without effect and asked why, that Claude refused to guide him further, and that he then turned to Grok and ChatGPT; the FIR says he sought information about the toxin abrin. Anthropic reported the activity to the FBI, whose information reached the NCCIA through Pakistan's foreign ministry on 22 September. Officers traced the teenager, seized his phone and chemicals, and took him into custody. The News reports that a court later granted him bail after his father forgave him and that he denied any plan. The allegations are untested.

Medium Unidentified facial recognition system

Reno, Nevada: a casino's facial recognition system wrongly matched a patron to a barred man in September 2023, and police arrested him and held him for about eleven hours until fingerprints confirmed his identity and he was released, his federal lawsuit alleges

Jason Killinger, a UPS truck driver, alleges in a federal civil-rights complaint filed in July 2025 that on 17 September 2023 the Peppermill Casino in Reno reported to police that its AI facial recognition software had identified him as a man barred from the casino months earlier. According to the complaint, casino security handcuffed him, and the responding Reno police officer, told the system had found a 100 percent match, did not examine his other identification and took him to jail as an unidentified person, where fingerprints confirmed who he was. The complaint says he was handcuffed for about four hours, starting in the casino security office, spent about eleven hours in police and jail custody, and was treated the next day for bruised wrists. According to the complaint, he was then charged with trespassing, and a prosecutor dismissed the charge without prejudice at the bench trial. Casino.org reports that he settled a suit against the casino. In March 2026 the court let him add the City of Reno and a municipal liability claim over facial recognition policies, but refused a new wrongful arrest claim against the officer on qualified immunity grounds. In September 2026 the ACLU, the ACLU of Nevada and the Innocence Project filed a proposed amicus brief supporting him. The case is ongoing.

Medium Unidentified image and video tool

India: police in Ujjain said AI-made videos falsely showing the whole Shahi Masjid being demolished helped draw a crowd that clashed with police on 28 September 2026; 15 people were taken into custody and influencers were booked

On 28 September 2026, as part of Ujjain's Shahi Masjid was being removed for a road-widening project, a crowd clashed with police; stones were allegedly thrown at security personnel and police used tear gas and a lathi charge. Ujjain Superintendent of Police Pradeep Sharma said videos made with AI, showing the entire mosque being demolished, and other provocative posts on social media had helped gather the crowd. Police said 15 people were taken into custody and reports put the total at eight FIRs at three police stations; social media influencers were booked for spreading the videos, and police said more than 100 pieces of content were taken down. Reports conflict on injuries: one outlet reported police saying six policemen were hurt, another reported the SP saying no one was injured.

If you or someone you know is struggling, free and confidential support is available. Find a helpline near you at Signpost.

Back to all incidents