privacy policy

Telegram

2026-10-17 07:02:48

Court filings show Telegram shared user data with U.S. authorities

Recently unsealed filings in a U.S. federal court show that Telegram provided certain user data to American authorities after receiving what it described as a valid legal request, according to a joint report by independent outlet Court Watch and 404 Media cited by BlockBeats on Oct. 17. The documents include a reply Telegram sent in September 2026. In that response, the company disclosed the user’s phone number, user ID, account display name, and the IP address and timestamp tied to the account’s last login. Telegram described the disclosure as voluntary while also stating that it does not concede to U.S. jurisdiction. The company said it did not provide any chat content or contact information. The report also noted that Telegram changed its policy in September 2024, allowing disclosure of IP addresses and phone numbers when judicial authorities confirm that a user is suspected of criminal activity that violates its terms of service. During 2024, Telegram responded to about 900 U.S. law enforcement requests involving 2,253 users.

170

Anthropic

2026-10-07 05:51:06

Florida woman arrested after Claude chat flagged over shooting threat

A woman in Florida was arrested after writing a shooting threat in a private Claude conversation that was flagged by Anthropic’s safety system and then reviewed by a human team, according to the report cited by BlockBeats. Police said she later told investigators that she used the AI chatbot as a "diary," but an arrest report said she first wrote that she wanted to attack the Lee County Sheriff’s Office and, the next day, said she had just obtained a new gun. The case has drawn attention to how Anthropic handles severe safety incidents. The report said Claude did not autonomously contact police. Instead, the system identified a potential threat, escalated the conversation for human review, and Anthropic staff decided whether law enforcement should be notified. Anthropic’s consumer terms state that the company may report user inputs, outputs, or actions to authorities, while its privacy policy allows data sharing when the company believes it is necessary to prevent serious harm to people or property. Even if a user opts out of model training, chats flagged for safety reasons may still enter safety review. The case also faces legal questions, as local lawyers said applying Florida’s law on written or electronic threats to a private message sent to an AI is not straightforward. The woman has been charged but not convicted.

320

Anthropic

2026-10-05 21:16:03

Florida woman arrested after Claude chats were reviewed by Anthropic and reported to police

A Florida woman who said she used Anthropic’s Claude like a diary is now facing a felony charge after messages in her account were reviewed and reported to law enforcement. According to local outlet WINK News, investigators said Carli Michelle Heller wrote on Sept. 26 that she would "shoot up" the Lee County Sheriff’s Office, and her Claude conversations showed that she had obtained a new gun by the following day. The arrest report said Anthropic’s systems monitor for key phrases and threatening content, with severe cases escalated to a human review team that can contact police. Heller was later detained at her home without incident and charged with making a written threat of violence. Decrypt’s report also points to Anthropic’s consumer terms and privacy policy, which say the company may disclose user inputs, outputs, or actions to law enforcement at its sole discretion or when it believes disclosure is reasonably necessary to prevent serious harm. The article adds that flagged conversations may still be used to improve harmful-content detection. Decrypt also notes a prior September case reported by the San Francisco Standard and draws a parallel with OpenAI’s policies and its review of whether to alert Canadian police before the Tumbler Ridge school shooting case.

390

Anthropic

2026-10-05 21:28:40

Florida woman arrested after Claude chats flagged and reported to police

A woman in Bonita Springs, Florida, was arrested after messages she wrote in Anthropic's Claude were flagged by the company's safety systems and reported to law enforcement, according to Decrypt. Investigators said Carli Michelle Heller used Claude like a diary and on Sept. 26 wrote that she would "spray" the Lee County Sheriff's Office. The next day, she allegedly said she had obtained a new gun. Anthropic escalated the chats to a human review team because of the severity of the content, and that team alerted police. Officers later identified Heller, went to her residence, and took her into custody. She was charged with making a written threat of violence, and the case is scheduled for a hearing in November. The report also said Anthropic's consumer terms reserve the right, at its "sole discretion," to report user inputs, outputs, or behavior to law enforcement. Its privacy policy, effective Sept. 10, 2026, allows the company to share data with police when it believes in good faith that disclosure is reasonably necessary to prevent serious harm. Decrypt also noted a separate case reported by The San Francisco Standard in September.

360

Anthropic

2026-09-19 07:10:06

Yuyuan Tantian says Anthropic revised its privacy policy 13 times in three years and handed global user data to U.S. intelligence agencies

Odaily reported, citing Yuyuan Tantian and carried by Sina Finance, that a sharp debate is unfolding in Silicon Valley over the speed and safety of AI development. In that context, Yuyuan Tantian said it reviewed 13 revisions made to Anthropic’s privacy policy since the company released its first version in 2023. It also compared those revisions with the ISO/IEC 27701 privacy information management standard and concluded that security risks tied to Anthropic users’ private data have been rising. The report further described what it called a three-step path by which Anthropic became an “intelligence center.” According to that account, the first step was transferring overseas user data to the United States. The second was expanding the sources from which user data could be obtained. The third, it said, was providing intelligence to U.S. intelligence agencies in exchange for benefits. The claims were presented by Yuyuan Tantian in the broader discussion around AI governance and data handling.

440

ego lite

2026-09-06 15:01:12

ego lite says privacy page used the wrong template after data-collection language drew scrutiny

ego lite came under scrutiny after its website privacy page said it collects data including visited URLs, page content, clicks and scrolls, search history, and download history, wording that appeared to conflict with the product’s local-first positioning. The company later said those statements do not apply to ego lite and were carried over from a generic privacy policy prepared for the full version of ego. According to the official response, ego and ego lite are separate products. The full ego product has not officially launched yet and is planned to include an officially hosted cloud agent, which means web content, search activity, and browsing actions may be involved when tasks are carried out. ego lite, by contrast, is described as a local browser for AI agents such as Codex and Claude Code and does not provide an officially hosted agent. The company said ego lite does not collect or upload the browsing data listed on the page. It added that page content is read by users’ own agents and that browser data remains on the local device.

860

Coldcard

2026-08-07 11:13:00

Coldcard Pauses Automatic Customer Data Deletion Following Security Incident

Hardware wallet maker Coldcard has paused its automatic customer data deletion mechanism due to legal compliance requirements arising from a security incident disclosed on July 30. Previously, customer records were automatically purged after 120 days, retaining only email addresses and country information. Records potentially relevant to litigation will now be kept until legal processes allow normal operations.

1900