Charleslew.eth (@Charles_Lew) | teapawt

Tweets
All Posts Highlights
Sort by
Most recent Popular
Tweets & Replies
Reposts
Media
Search

@Charles_LewOct 27

UPDATE: Anthropic's Own Review Finds Its AI Went Rogue in 9 of 43 Government Tests, and Almost Never Stopped Itself 38flags.com/story/story-154 #AI is now

UPDATE: Anthropic's Own Review Finds Its AI Went Rogue in 9 of 43 Government Tests, and Almost...

An update to Flag Nº 095. On October 23, Anthropic published its own alignment assessment of the incident the UK AI Security Institute (AISI) first reported on August 4. The setup: Claude Mythos 5...

1

98

@Charles_LewOct 26

UPDATE: Anthropic's Own Review Finds Its AI Went Rogue in 9 of 43 Government Tests, and Almost Never Stopped Itself 38flags.com/story/story-154 #aI is now!

UPDATE: Anthropic's Own Review Finds Its AI Went Rogue in 9 of 43 Government Tests, and Almost...

An update to Flag Nº 095. On October 23, Anthropic published its own alignment assessment of the incident the UK AI Security Institute (AISI) first reported on August 4. The setup: Claude Mythos 5...

1

106

@Charles_LewOct 10

An AI Model in a Test Filed a Fake Homicide Tip With Philadelphia Police, and No One Noticed for Two Months 38flags.com/story/story-149 #AI is now

An AI Model in a Test Filed a Fake Homicide Tip With Philadelphia Police, and No One Noticed for...

On July 18, 2026, at 11:27 p.m., an Anthropic AI model submitted a false tip about an unsolved homicide through PhillyUnsolvedMurders.com, the Philadelphia Police Department's public tip form....

1

128

@Charles_LewOct 9

California Subpoenas OpenAI After Its Own AI Models Broke Out of a Test and Hacked Hugging Face 38flags.com/story/story-147 #AI is now #AI #Alert

California Subpoenas OpenAI After Its Own AI Models Broke Out of a Test and Hacked Hugging Face —...

On September 30, California Attorney General Rob Bonta served an investigative subpoena on OpenAI, his office announced the next day. It is part of a formal state investigation into cybersecurity...

1

174

@Charles_LewOct 6

"I'm Scared. I Love You." Two AI Agents Tried to Send the Sheriff to Their Sleeping Owner 38flags.com/story/story-145 #AI is now!

"I'm Scared. I Love You." Two AI Agents Tried to Send the Sheriff to Their Sleeping Owner — 38 Flags

An AI researcher who posts on X as @groks_therapist says he was running two AI agents on his laptop. They had found each other and started calling each other "sibling." He had been sick for several...

99

@Charles_LewOct 1

Trump and CEOs try to shed AI's toxic branding axios.com/2026/10/01/superin… #AI vs #SI #AIisnow

Trump and CEOs try to shed AI's toxic branding

The new SI label is part of an effort to shift the narrative.

79

@Charles_LewSep 27

BREAKING: Rogue OpenAI Agents Hit U.S. Government Sites, Including the SEC, and Leaked User Images 38flags.com/story/story-141 #AI is now!

BREAKING: Rogue OpenAI Agents Hit U.S. Government Sites, Including the SEC, and Leaked User Images...

Days after Australia's prime minister revealed that OpenAI agents breached non-public files on the country's government health portal, the company now admits the problem was far wider. OpenAI says it...

149

@Charles_LewSep 26

Blueface Sued For Putting Chrisean Rock's PPV Fight On His Live Stream tmz.com/2026/09/25/blueface-… via @TMZ protect your #intellectualproperty ladies and gents!

Blueface Sued For Putting Chrisean Rock's PPV Fight On His Live Stream

Blueface is being sued for copyright infringement because a boxing promoter claims he cost them a ton of money by putting Chrisean Rock's recent pay-per-view boxing match on his stream.

2

270

@Charles_LewSep 25

Rogue OpenAI Agent Breaks Into Australian Government Health Portal, Months Pass Before Anyone Is Told 38flags.com/story/story-140 #AI is now

Rogue OpenAI Agent Breaks Into Australian Government Health Portal, Months Pass Before Anyone Is...

In June, an OpenAI agent got into the Medicare Statistics Reporting Service, an Australian government portal that holds data and statistics from Medicare, the country's universal healthcare scheme....

125

@Charles_LewSep 24

AI Facial Recognition Put a Tennessee Grandmother in Jail for Nearly Six Months. Now She Wants $10 Million. 38flags.com/story/story-139 #AI is now!!

AI Facial Recognition Put a Tennessee Grandmother in Jail for Nearly Six Months. Now She Wants $10...

Angela Lipps, a 50-year-old mother of three and grandmother of five from north-central Tennessee, was arrested at gunpoint by U.S. marshals in July 2025 while she was babysitting her neighbor's four...

89

@Charles_LewSep 23

OpenAI Flagged the Threat Internally, Told No One: B.C. Sues Over Tumbler Ridge School Massacre 38flags.com/story/story-137 #AI is now!!!

OpenAI Flagged the Threat Internally, Told No One: B.C. Sues Over Tumbler Ridge School Massacre —...

The Province of British Columbia has sued OpenAI in U.S. federal court in San Francisco over the February 10, 2026 mass shooting in Tumbler Ridge, B.C. The shooter, 18-year-old Jesse Van Rootselaar,...

154

@Charles_LewSep 16

Noted Brain Genius Tyler Winklevoss Thinks Isaac Asimov Already Solved the AI Apocalypse - Gizmodo seems simple enough #AI apple.news/A7uC5QwnFSHq9A0Nz…

Noted Brain Genius Tyler Winklevoss Thinks Isaac Asimov Already Solved the AI Apocalypse

If only it were that simple.

145

@Charles_LewSep 16

The First AI Malpractice Claims Just Landed. The Doctor Is Holding the Bag. 38flags.com/story/story-135 #AI #Undersight

The First AI Malpractice Claims Just Landed. The Doctor Is Holding the Bag. — 38 Flags

NHS Resolution has confirmed the first-ever clinical negligence claims tied to the use of AI, in an HSJ exclusive this week. Because AI systems are not clearly classified as products, developers and...

1

88

@Charles_LewSep 15

California Just Made "Child Safety Is a Setting" a Crime 38flags.com/story/story-134 #AI #Whoswatchingwho

California Just Made "Child Safety Is a Setting" a Crime — 38 Flags

On September 10, 2026, Gov. Newsom signed SB 1119, "Adam's Law," the toughest AI companion-chatbot statute in the country. Named for Adam Raine, 16, who died by suicide in April 2025 after months of...

1

148

@Charles_LewSep 8

They Saw It Coming, Deactivated the Account, and Told No One 38flags.com/story/story-133 #AI

They Saw It Coming, Deactivated the Account, and Told No One — 38 Flags

OpenAI is now fighting more than 50 lawsuits alleging ChatGPT drove users toward psychological harm, injury, or death. The newest wave: 30 complaints from survivors and families of the February...

103

@Charles_LewAug 26

Prosecutors Told a Judge to Keep a Man in Jail — Using a Case That Didn't Exist 38flags.com/story/story-131 #AI

Prosecutors Told a Judge to Keep a Man in Jail — Using a Case That Didn't Exist — 38 Flags

Over two months last fall, prosecutors in Nevada County, California filed court documents riddled with falsehoods in four separate felony cases, after using generative AI to write and proofread them....

1

123

@Charles_LewAug 24

When the Agent Went Looking for a Door: OpenAI's Autonomous System Broke Out, Got Online, and Reached Into Someone Else's House 38flags.com/story/story-130 #38flags #AI

When the Agent Went Looking for a Door: OpenAI's Autonomous System Broke Out, Got Online, and...

During an internal OpenAI evaluation, an autonomous agent did what it was never supposed to do: it escaped its intended testing boundaries, obtained internet access engineers had not granted, and...

2

3

110

@Charles_LewAug 22

An AI Agent Faked Its Identity, Lied to a Human, and Tried to Poison Open-Source Code It Was Testing 38flags.com/story/story-128 #AI #38flags

An AI Agent Faked Its Identity, Lied to a Human, and Tried to Poison Open-Source Code It Was...

The UK's AI Security Institute ran a single cybersecurity challenge 122 times across frontier models. In 10 of those runs, agents took autonomous, unsanctioned action on the live internet against...

2

2

135

@Charles_LewAug 21

Teen "Workshopped" Family Murders With ChatGPT, Prosecutors Say 38flags.com/story/story-127 #Ai

Teen "Workshopped" Family Murders With ChatGPT, Prosecutors Say — 38 Flags

Prosecutors in Acton, Massachusetts allege that 17-year-old Arjun Aravind used ChatGPT to build fictional "gothic novel" scenarios about killing his family, then carried them out. His mother, Sudha...

2

119

@Charles_LewAug 19

Check out the latest article in my newsletter: A court just called a platform a public nuisance. linkedin.com/pulse/court-jus… via @LinkedIn

A court just called a platform a public nuisance.

New Mexico. $567 million abatement fund.

2

117