Foot off the gas? Never

Trump's message to frontier AI labs this weekend was to keep accelerating. His refusal to regulate lands hardest on publishers trying to make those same labs pay for training data.

Guardrails, but not really

Asked about safety limits, Trump waved the idea off and blamed critics for inventing problems that will never arrive.

“We could put guardrails. We can do this and that.”

The Hoax Buster posts

On Truth Social he lumped AI risk in with the Russia investigation and climate change, casting himself as the man debunking it.

“I am the Hoax Buster, and I'm right now breaking another Hoax”

Beat China at all costs

He told labs to ignore calls for a slowdown, pointed at Xi's hands-off approach, and complained that permitting pushed a Google plant toward Finland.

“AI, and Data Centers, will be the Greatest Economic Development Engine in History”

Bad math for the lawsuits

Publishers are suing OpenAI, Google, Microsoft, and Anthropic over training on their journalism. A White House fixated on winning the AI race has little reason to bless a huge copyright payout.

What the newsrooms claim

News outlets say their work was scraped to power answers and knowledge panels with no payment. The New York Times sued OpenAI and Microsoft, then said it found substantial evidence of stolen data.

The DOJ picks a side

This month the Justice Department told a New York court that training models on copyrighted material is fair use, and that a narrower reading would hold back American progress.

“Constraining LLM development under a misunderstanding of fair use doctrine would thwart...creative and scientific progress”

Where exactly is the difficulty

Ziff Davis CEO Vivek Shah, whose company sued OpenAI over copied PCMag articles, noticed the DOJ admitting in a footnote that licensing deals already exist and work.

“So, where exactly is the difficulty? What onerous obligation are we placing on frontier labs that would slow their progress?”

A rounding error on $750B

OpenAI expects to spend roughly $750 billion on compute by 2030. The US music industry collects about $20 billion a year in royalties. Shah's point: paying news would barely register.

“it would represent a rounding error to LLM companies' overall expenses, but be meaningful for content producers of all sizes”

Blue links kept the lights on

Search scraped the web too, but it sent readers back. That traffic funded the whole digital pivot publishers were forced into.

Chatbots end the click

AI answers cite sources, yet almost nobody taps through once the summary has told them enough. Traffic falls, and revenue follows it down.

The model eats itself

Brookings fellow Courtney Radsch calls AI platforms the new tollbooths. Starve the newsrooms and you starve the training data, sliding toward a dead internet of bots recycling bots.

“AI systems depend on a continuous supply of high-quality human content to remain useful”

Deals for the big names only

The Washington Post signed with OpenAI, the AP with Microsoft. Smaller outlets lack the leverage to negotiate anything, and neither a landmark ruling nor a act of Congress looks close.

Paywalls are the fallback

Without a licensing framework, Shah expects publishers to wall off their work and charge for it directly. Readers pay more, the free web shrinks, and writers lose ways to earn.

“publishers will increasingly erect barriers and paywalls to restrict their own content and extract those fees”