Blog

Elon Musk’s social media platform X said it will suspend creators from its revenue-sharing program if they post AI-generated videos depicting armed conflict without clearly disclosing that the footage was created using artificial intelligence.

In a post on Tuesday, X’s head of product Nikita Bier said the company is revising its Creator Revenue Sharing policies to maintain authenticity on the platform’s timeline and “prevent manipulation of the program.”

“During times of war, it is critical that people have access to authentic information on the ground,” Bier wrote. “With today’s AI technologies, it is trivial to create content that can mislead people.”

Creators who violate the rule will lose access to the platform’s Creator Revenue Sharing program for 90 days, Bier wrote. Repeat violations will lead to permanent removal from the monetization program.

The policy change comes as AI-generated videos claiming to show scenes of escalating violence in the Middle East following missile strikes by the U.S., Israel, and Iran last week.

On Monday, an AI-generated clip on X showing an airstrike on the Burj Khalifa in Dubai was viewed over 8 million times; at the same time, another version of the clip was viewed over 42,000 times on Instagram.

The United Nations has warned that deepfakes and AI-generated media threaten information integrity, particularly in conflict zones where fabricated images or videos can spread hate or misinformation at scale.

This concern was realized during Russia’s invasion of Ukraine, a deepfake video circulated online appearing to show Ukrainian President Volodymyr Zelensky urging Ukrainian troops to surrender. Officials quickly debunked the video, and Zelensky later released a message rejecting the claim.

According to Bier, enforcement will rely on several signals, including posts that receive a Community Note identifying the video as AI-generated, along with metadata or other indicators suggesting the footage was produced using generative AI tools.

By tying enforcement to monetization, X’s policy focuses specifically on the financial incentives creators have to post fake videos that drive clicks and views.

“We will continue to refine our policies and product to ensure X can be trusted during these critical moments,” Bier wrote.

The Supreme Court of Colombia denied a cassation appeal, arguing that it was generated by AI. But the same tool the court used to determine the appeal’s purported AI origins said that its own ruling also received generative help.

Is it a double standard by the court, or faulty tools at play?

“Faced with a well-founded suspicion that the brief submitted by the attorney had not been drafted by the legal professional himself, the court submitted the text to the Winston AI tool,” the court argued. “Its analysis indicated that the document contained only 7% human content, evidencing a marked influence of automated writing and leading to the conclusion that it had been produced using artificial intelligence.”

After running the analysis with other tools that provided similar results, the court ruled that “since the filing cannot be regarded as a duly submitted pleading, its dismissal as inadmissible is required.”

But when the court’s ruling faced similar scrutiny from legal experts, it showed similar results.

“I submitted the text of Auto AP760/2026 from the Supreme Court to the same Winston AI software cited in the ruling,” attorney Emmanuel Alessio Velasquez wrote on X on Tuesday. “The result: The document contains 93% AI-generated text.”

“If the very ruling that condemns the use of artificial intelligence scores that percentage, the methodological fragility of using these detectors as argumentative support becomes self-evident,” he argued in a subsequent tweet.

Within hours of the court posting a thread about the decision on X, lawyers began running their own tests. Velasquez’s post went viral in legal circles, accumulating tens of thousands of views.

We ran the test on the court’s verdict, as well, and things initially didn’t look great. When GPTZero scanned only the opening words of the court text, it returned a 100% AI result.

When the same tool processed a longer version including the factual background section, it reversed course entirely: 100% human.

The tool is simply not reliable enough to be trusted in court or in situations that would require a high degree of certainty.

Colombian attorneys reacted quickly with their own experiments. Criminal defense lawyer and lecturer Andres F. Arango G, submitted a court filing from 2019, years before the large language models these tools were trained to detect even existed, and it came back claiming 95% AI generation.

“These tools then invite you to ‘humanize’ the article through their paid services,” he wrote on X, noting an obvious commercial incentive baked into the detection business model.

Nicolas Buelvas ran his 2020 undergraduate thesis on the principle of trust in criminal law. The result? 100% AI.

Dario Cabrera Montealegre, another Colombian attorney, pointed out the hypocrisy of relying on technology to try to combat it.

“The court is using AI to determine if there was AI,” he said. “Something contradictory from my practical point of view.”

Beyond legal circles, further tech-savvy individuals pointed out the dangers of excessive reliance on AI flagging tools.

“To date, there is no publicly accessible tool that can accurately define the percentage of AI use when drafting a text,” Carlos Alejandro Torres Pinedo argued. “What is worse: No one can publicly verify the source code behind these detection platforms. How can they be used to delegitimize someone’s right of access to justice?”

The technical reasons for these failures are well-documented. AI detectors measure statistical patterns: sentence length, vocabulary predictability, and a quality that researchers call “burstiness,” which refers to the natural rhythm variation humans introduce in their writing.

The problem is that formal legal prose, academic writing, and texts produced by people who write in a second language share many of those same statistical signatures.

Studies on AI detection

A 2023 study published in Patterns found that more than 61% of Test of English as a Foreign Language (TOEFL) essays by non-native English speakers were incorrectly flagged as AI-generated.

A systematic review by Weber-Wulff that same year concluded no available tool is either precise or reliable. Turnitin acknowledged in June 2023 that its own detector produced higher false positive rates when the AI content level in a document fell below 20%.

Even OpenAI had to take down its own AI detection tool following constant inaccuracies and an inability to do its actual job.

Universities have been grappling with this for years. Vanderbilt disabled Turnitin’s AI detector in 2023 after estimating it would generate around 3,000 false positives annually.

The University of Arizona dropped AI-detection features from its plagiarism software after a student lost 20% of a grade on a false positive. A 2024 case at UC Davis saw 17 linguistics students flagged, 15 of them non-native English speakers.

The pattern is consistent. The tools penalize the people who write most formally, most repetitively, or most carefully, exactly the profile that lawyers, academics, and second-language speakers fit.

The cultural fallout has bordered on absurdity. Across writing and journalism circles, people have started avoiding em dashes in their work, not because of any style guide, but because AI language models use them frequently and detection tools (and people) have taken notice.

Writers are self-editing natural punctuation out of fear of algorithmic suspicion. Beyond the written world, artists have suffered the wrath of moderators and colleagues for making art pieces that look AI

Colombia’s two rulings—AC739-2026, in which the Civil Chamber fined a lawyer for citing 10 nonexistent AI-generated precedents in February, and AP760-2026—are emerging as some of the region’s first judicial decisions directly confronting the misuse of generative AI in legal filings.

Colombia’s judicial branch adopted formal guidelines in December 2024 that regulate how judges and court staff can use artificial intelligence.

The rules allow AI to be used freely for administrative and support tasks, such as drafting emails, organizing agendas, translating documents, or summarizing texts, while permitting more sensitive uses, like legal research or drafting procedural documents, only with careful human review.

The guidelines explicitly prohibit relying on AI to evaluate evidence, interpret the law, or make judicial decisions, emphasizing that human judges remain fully responsible for all rulings and must disclose when AI tools were used in preparing judicial materials.

These guidelines, compiled in the “PCSJA24-12243” agreement, could be used to contest such a decision.

The Supreme Court has not yet issued any additional statement in response to the backlash over its choice of detection tools. The ruling didn’t have em dashes, either.

OpenAI on Tuesday announced the rollout of GPT-5.3 Instant, an update to ChatGPT’s default model aimed at making conversations feel less awkward and preachy, and “more directly” helpful.

In a product post, OpenAI said the new version reduces overly cautious refusals, trims unnecessary disclaimers, and delivers more accurate answers. The changes reflect user complaints that earlier versions could sound stiff or overbearing in everyday interactions.

“More accurate, less cringe,” OpenAI wrote on X. “We heard your feedback loud and clear.”

Rather than introducing new capabilities, the update targets routine interactions.

“Part of the day-to-day experience with ChatGPT comes down to interacting with the model,” an OpenAI spokesperson told Decrypt. “This update focuses on addressing common user feedback we’ve received, including reducing unnecessary refusals, cutting down on caveats, and making answers more direct and useful in everyday conversations.”

OpenAI said earlier versions sometimes declined questions they could safely answer, or interrupted responses with lengthy explanations about safety limits.

“GPT‑5.2 Instant eventually answers the question, but in an attempt to explain its safety boundaries, leads with a lengthy preamble about what it cannot help with,” OpenAI wrote. “GPT‑5.3 Instant, on the other hand, gets right into the response.”

OpenAI reported improvements in factual reliability alongside the tone changes, claiming that internal evaluations showed hallucination rates dropped by nearly 30%.

“On the higher-stakes evaluation, GPT‑5.3 Instant reduces hallucination rates by 26.8% when using the web and 19.7% when relying only on its internal knowledge, compared to prior models,” OpenAI said. “On the user-feedback evaluation, hallucinations decrease by 22.5% with web use and 9.6% without web access.”

OpenAI did not explain what it defines as “cringe,” but noted that the new model includes stronger writing abilities, comparing GPT-5.2 and 5.3’s ability to write poetry.

“5.4 sooner than you think,” the company said in a separate post, which drew swift mockery from users on X, suggesting that the tease was due to recent backlash against the firm for its deal with the Pentagon.

GPT-5.3 Instant replaces the default ChatGPT model starting today, the company said, while GPT-5.2 Instant remains accessible under legacy options for paid subscribers during a transition period ending in early June.

The update drew mixed reactions on social media. Some users praised the focus on more direct responses without unnecessary disclaimers, while others argued the real “cringe” at play was agreeing to a contract with the U.S. Department of Defense when rival Anthropic declined due to safety concerns.

Others complained that GPT-5.3 would never match the sense of intimacy many associated with the now-depreciated GPT-4o, and called for the popular model’s return.

Last summer, OpenAI faced a surge in backlash after the company abruptly replaced the popular GPT-4o with GPT-5, prompting complaints that the new model felt colder and less supportive. Users flooded forums with criticism, and some threatened to cancel subscriptions, leading OpenAI to restore GPT-4o for paid users.

In January, OpenAI announced that GPT-4o and its variants would be officially retired as of February 13.

“I think we’ve learned a lesson about what it means to upgrade a product for hundreds of millions of people in one day,” OpenAI CEO Sam Altman said at the time, calling the reversal a wake-up call and “a lesson in upgrading a product used by hundreds of millions of people” at once.

Editor’s note: This story was updated after publication to include comments from OpenAI.

Police in South Korea say people are using crypto to pay for intimidation tactics that include vandalizing front doors, leaving threatening messages, and spreading human waste, according to a local news report from Hankyoreh. 

The so-called “private revenge” attacks were reportedly ordered through social messaging app Telegram, with alleged perpetrators paid between $337-$675 or 500,000-1,000,000 South Korean won worth of cryptocurrency—though different suspects in at least three cases spanning back to December claim they do not know who ultimately paid them.

In the latest two events, which have occurred in the last week, individuals identified as “Mr. Lim” and “Mr. K” by the report were arrested and charged by South Korean police. The pair both vandalized the front door of residences and were accused of dropping defamatory leaflets, at least some of which contained the message “I will not leave you alone.”

In Mr. Lim’s case, the man, identified to be in his 20s, also scattered food waste and was accused of spreading human waste on a nearby stairwell. The attacks occurred in the Suwon District of South Korea, outside Seoul.

“Police believe that the individuals arrested this time committed the crimes under the direction of a private revenge organization operating on Telegram, and are tracking down their superiors,” the local report reads. 

The recent cases are also under investigation regarding their potential connection to a December 7 vandalization, which also saw the dropping of defamatory leaflets and payments to three individuals via cryptocurrency.

The crypto crime spree follows a recent Bitcoin dispute in South Korea that led to attempted murder charges. According to authorities in that case, an individual laced his business partner’s coffee with methomyl—a banned and toxic insecticide—after the colleague allegedly mismanaged his Bitcoin investments.

The country has also seen regulators recently face scrutiny regarding their inability to find an internal system flaw in crypto exchange Bithumb, which led to the erroneous distribution of up to 2,000 BTC or $137 million to hundreds of customers, instead of 2,000 won ($1.35). All told, Bithumb credited users with $43 billion in Bitcoin, though it realized the mistake within minutes and clawed back most of those funds.

Furthermore, it was recently discovered that police officers from Gangnam Police Station had lost access to $1.4 million worth of Bitcoin more than four years ago. Plus, the nation’s tax service (NTS) publicly shared the seed phrase for three crypto wallets that held $4.8 million worth of tokens at face value in a press release.