Poynter: An FT column was condensed with AI before submission
In August 2026, Harvard professor Ricardo Hausmann submitted an opinion column to the Financial Times that had been shortened using an AI tool before submission. The FT discovered this during the editorial process and appended a note to the published piece stating that AI was used to condense a longer draft prior to submission.
The Financial Times has a published AI policy stating that journalism must be “reported, written and created by our journalists and editors.” The policy allows AI to assist with specific tasks like story discovery and accuracy checks, but prohibits AI participation in the writing and production process itself. Hausmann, as an outside contributor, was not a staff journalist — and the FT’s disclosure suggests the organization’s policy had not been communicated clearly to external columnists.
The incident matters as a case study because no factual fabrication occurred. Hausmann wrote the column, developed the argument, and conducted the underlying work. The AI tool performed a single mechanical task: shortening an already-written draft. This is where the article raises the question it cannot fully resolve. Where exactly is the line between a permitted editing aid and prohibited AI assistance? If using AI to trim paragraphs violates policy, does that mean spell-check does too? What distinguishes acceptable from unacceptable AI use when the same tool can perform either function depending on how it is used?
The case also reveals a structural gap in how news organizations manage AI restrictions. Most AI policies are written for staff journalists working within a newsroom. Outside contributors — freelancers, academics, industry columnists — operate without the same onboarding and are often unaware of the specific rules. As AI becomes embedded in general-purpose writing software, publications will need to communicate their restrictions explicitly at the point of submission, not assume outside contributors have read internal policy documents.
For editorial teams: this case provides a concrete prompt for reviewing contributor agreements and submission guidelines. For writers working across multiple outlets: the article illustrates that AI policies vary, are inconsistently communicated, and are becoming a factor in how work is received and labeled.