Over 100 fake citations slip through peer review at top AI conference
An analysis of nearly 5,000 accepted research papers from the AI conference NeurIPS 2025 uncovered more than 100 fabricated citations – despite review by multiple experts.
Read full article about: South Koreans now spend more on AI subscriptions than Netflix each month
South Koreans now spend more per month on AI subscriptions than on Netflix. According to Hankyung Aicel, payments for seven AI services, including ChatGPT and Gemini, hit an estimated 80.3 billion won (roughly $55-60 million) in December 2025. That's more than the average monthly Netflix subscription revenue in Korea during 2024, which came in at 75 billion won (around $50-55 million). One important caveat: the AI figure includes business payments, while Netflix is a consumer-only service.
Credit card payments for AI services jumped from 52,000 transactions in January 2024 to 1,666 million in December 2025. Private customers paid an average of 34,700 won (about $24), while businesses spent 107,400 won (roughly $74). ChatGPT dominated with 71.5 percent of all payments, followed by Gemini at 11.0 percent and Claude at 10.7 percent. According to Hankyung Aicel CEO Kim Hyung-min, Korea's subscription market continues to grow, and generative AI is becoming a regular subscription product.
For context: Netflix reports revenue per subscription of around $7 for Asia-Pacific, compared to roughly $17 in the US and Canada. That's significantly higher revenue per subscription per month.
Read full article about: Bandcamp bans AI-generated music
Music platform Bandcamp now prohibits music created entirely or substantially by generative AI. The company says the new policy protects human creativity and the direct connection between artists and fans. The updated rules also strictly ban using AI tools to imitate specific artists or styles.
Unlike most streaming services, Bandcamp focuses on direct purchases of music and merchandise, letting fans support creators financially without intermediaries.
Users can now report content that sounds heavily AI-generated. Bandcamp reserves the right to remove music from the platform based on suspected AI origins alone.
Read full article about: Microsoft pledges to cover data center power costs as community pushback grows
Microsoft is rolling out a new initiative for AI data centers after facing mounting opposition from communities across the US. The company says it will fully cover the power costs of its data centers, ensuring residents won't see higher electricity bills as a result. The announcement comes as data center regions like Virginia, Illinois, and Ohio have seen electricity prices climb 12-16 percent faster than the national average.
Beyond power costs, Microsoft is making several other commitments: the company will stop requesting local tax breaks, cut water consumption by 40 percent by 2030, and replenish more water than it uses. Microsoft President Brad Smith told GeekWire that the industry used to operate differently and now needs to change its approach. Trump previewed the announcement on Truth Social before Microsoft made it official.
As part of the initiative, Microsoft also plans to train local workers and invest in AI education programs in affected communities.
Read full article about: AI models don't have a unified "self" - and that's not a bug
Expecting internal coherence from language models means asking the wrong question, according to an Anthropic researcher.
"Why does page five of a book say that the best food is pizza and page 17 says the best food is pasta? What does the book really think? And you're like: 'It's a book!'", explains Josh Batson, research scientist at Anthropic, in MIT Technology Review.
The analogy comes from experiments on how AI models process facts internally. Anthropic discovered that Claude uses different mechanisms to know that bananas are yellow versus confirming that the statement "Bananas are yellow" is true. These mechanisms aren't connected to each other. When a model gives contradictory answers, it's drawing on different parts of itself - without any central authority coordinating them. "It might be like, you're talking to Claude and then it wanders off," says Batson. "And now you're not talking to Claude but something else."
The takeaway: Assuming language models have mental coherence like humans might be a fundamental category error.
Read full article about: Trump's AI plan could affect his own voters
Trump attempts to block state AI laws by withholding broadband billions, but faces shaky legal ground.
"I think the administration has a 30 to 35% chance of this working legally," says Dean Ball, a former White House official who contributed to the administration's AI Action Plan.
The executive order directs the Commerce Department to block states with onerous AI regulations from the $42 billion Broadband Equity, Access, and Deployment program (BEAD), reports Reuters in an analysis of the new order. However, experts doubt whether Congress intended to give the administration authority over state AI regulation when it authorized broadband funding. Furthermore, the move risks political blowback from within the party: Republican governors like Ron DeSantis have previously spoken against federal interference, and withholding funds would impact rural voters—a key demographic that supported Trump by wide margins.
