Seattle Police Communications Director Used Unapproved AI Tools for Interview Prep and Blog Posts
Barbara DeLollis, the Seattle Police Department's communications director, used ChatGPT and Perplexity-neither approved by the city-to draft interview scripts, rewrite published blog posts, and create communications strategy documents, according to records obtained through a public disclosure request.
The city's information technology department confirmed both tools are unauthorized for city employee use. DeLollis used ChatGPT to produce a sample blog post and Perplexity for other documents including a "Comprehensive Communications Toolkit for a Police Department Exiting a Consent Decree."
What the Records Show
DeLollis asked Perplexity to rewrite a blog post about a nuisance motel with this prompt: "hi make this a better story for the public of a city that doenst liek crime or disorder" (sic).
Other requests included help preparing a woman for her first media interviews. One prompt read: "So we know why woken over prepare for media interviews but for our client we want to frame this advice on a positive way to prevent them from feeling negative. Help."
The prompts themselves are riddled with typos and unclear phrasing, making them difficult to interpret. In one request about the department's Blotter blog, DeLollis wrote: "Lost thengoala for a daily police blog where uoupost both police actions and responses to 911 callbut also show in compelling ways the other impactand resorts of work that a cutting edge evidencebased department does."
The Department's Response
The Seattle Police Department said DeLollis used AI tools only "to evaluate their utility" in communications functions. An SPD spokesperson told reporters that "a department employee tested various AI tools to evaluate their utility in communication functions like editing, interview preparation, and blog strategy to see if they could offer fresh perspectives."
The Office of Police Accountability determined that using AI tools without acknowledgement violated city policy at the time and issued a supervisor action-essentially a reprimand. The department said it "does not condone using generative AI to write narratives or communications."
Incomplete Records Raise Questions
The SPD provided seven documents in response to the disclosure request, all produced by DeLollis. Two of those documents-guides about communicating the end of the federal consent decree-lack the written prompts that preceded them, an omission that raises concerns about whether the records are complete.
The department confirmed that Police Chief Shon Barnes has never used generative AI and that the seven documents represent every use of AI by DeLollis and the entire communications team.
An anonymous complaint filed with the Office of Police Accountability last year cited apparent AI use in other SPD communications, including a bio of Barnes' chief of staff and a bullet-pointed statement about a violence prevention effort. A blog post in that complaint contained passive-voice sentences such as "On Thursday, we were confronted with a targeted homicide occurring in front of a place of worship."
City Moves to Block Unapproved Tools
Mayor Katie Wilson's office said unapproved AI software is now blocked on city devices "to ensure compliance with critical privacy, transparency, and records protections." The mayor expects employees to use AI tools in compliance with city policy.
Washington's public disclosure law requires city employees to produce all records responsive to a request, including those created on personal devices or using personal email and phone numbers. It's unclear whether DeLollis created the chat prompts from a city computer or personal device.
What This Means for Communications Teams
AI for PR & Communications professionals face growing pressure to use generative tools for efficiency. This case illustrates the risks: unapproved tools can create compliance problems, and AI-generated content often needs substantial revision to meet professional standards.
For communications staff, the takeaway is straightforward. Verify that any AI tool your organization uses is approved by your IT department and compliant with local policy. Document your use clearly. And recognize that AI outputs-especially from tools like Perplexity that sometimes hallucinate sources or provide generic advice-often require significant editing to be useful.
Your membership also unlocks:
