A Financial Times editor’s note admitted artificial intelligence helped shape a high-profile op-ed attacking President Trump’s tariffs, igniting a credibility fight the paper tried to avoid.
Story Snapshot
- Financial Times disclosed artificial intelligence condensed a Harvard professor’s anti-tariff op-ed before editors reviewed it.
- The paper’s public policy pledges journalism is created by its reporters and editors, inviting questions about limits on tools.
- Social media backlash labeled the piece “AI slop,” citing an unverified detector score as proof.
- The episode spotlights growing distrust of elite institutions and confusion over what “AI use” means in media.
What the Financial Times Disclosed and Why It Matters
The Financial Times added an editor’s note to Ricardo Hausmann’s column. The note said artificial intelligence was used to condense a longer draft before he sent it to the paper and before editors worked on it. The paper has a public policy that promises journalism “will continue to be reported, written and created” by its staff. That pledge set a high bar. The disclosure raised a narrow but important question: when does assistive software cross into authorship?
The op-ed argued that President Trump’s tariff policy harms income the United States earns from knowledge deployed abroad. That policy debate is serious and technical. But the story shifted fast to process, not substance. Readers saw the admission about artificial intelligence and wondered who really wrote what. The gap between a strict-sounding policy and a fuzzy real-world workflow fueled doubts on both the left and the right.
How Backlash Turned a Policy Note Into a Public Pile-On
Coverage from major outlets captured a wave of mockery online. Users called the prose “the worst AI slop,” and some pointed to a screenshot of a detector that scored the column as mostly machine-written. A separate report said the detector estimate was 71 percent artificial intelligence, shared by a political scientist on social media. Those claims traveled far. The Financial Times did not endorse any detector result. The number functioned as a meme, not a verified audit.
Artificial intelligence detectors are often treated like breathalyzers. But experts warn many tools are not reliable proof of authorship. Studies show false positives and inconsistent results across detectors, especially with non-standard styles. That means a high score can mislead the public. In this case, the only confirmed fact is the paper’s note about condensation before submission. There is no published, independent analysis of drafts to show deeper machine rewriting.
Where Policy Meets Perception in a Distrustful Era
The Financial Times says it will use artificial intelligence to assist journalists and to make secondary content, with disclosure, while protecting core integrity. That stance mirrors many newsrooms. Still, the promise that staff “create” the journalism can clash with any revealed machine role. When the op-ed also attacks the sitting administration’s tariffs, some readers on the right see bias. Others on the left see elite gatekeeping. Both read the gap as another sign the system talks one way and acts another.
When the professor's paper gets called "slop." Harvard economist Ricardo Hausmann is facing scrutiny after AI use surfaced in his Financial Times column criticizing Donald Trump's tariffs. 📝 #Harvard #Hausmann #DonaldTrump #FinancialTimes #AISlop https://t.co/B9yOHvzwCM
— TAXI (@designtaxi) August 17, 2026
Harvard’s Growth Lab promoted the column’s thesis on tariffs and America’s overseas income stream, reinforcing the piece’s policy frame. Yet process overshadowed message because trust is fragile. People suspect rules are elastic for insiders. The fastest fix is sunlight. A clear timeline of drafts and tools would help. If the paper plans a review, publishing the boundaries between condensing, rewriting, and authorship would steady the standard without chilling careful opinion writing.
Sources:
townhall.com, nypost.com, growthlab.hks.harvard.edu, foxnews.com, aboutus.ft.com, ora.ox.ac.uk
