Recently, Microsoft made headlines for publishing — and then quickly retracting — an AI-generated article that suggested visitors to Ottawa should treat a local food bank as a tourist attraction. The piece titled “Traveling to Ottawa? Discover Essential Experiences You Can’t Miss!” included various recommendations, such as attending a baseball game, paying respects to fallen soldiers at a military museum, and, surprisingly, visiting the Ottawa Food Bank. Paris Marx initially highlighted this inappropriate suggestion on X (formerly known as Twitter), stating, “People who come to us have jobs and families to support, as well as expenses to pay. Life is already difficult enough. Consider going into it on an empty stomach.” This statement starkly emphasized the insensitivity of the article’s content.

Before the article was removed, it was featured on Microsoft Start, the company’s AI-driven news aggregation service that replaced Microsoft News back in 2021. Following a report from The Verge regarding the article’s troubling suggestion about visiting a food bank while hungry, Microsoft’s senior director, Jeff Jones, stated to the publication, “This article has been removed, and we are investigating how it made it through our review process.” This incident raises important questions about the oversight of AI-generated content and the standards for quality control.

Currently, the original URL of the article displays a message indicating, “This page no longer exists. A new search page will load automatically.” In response to the incident, The Verge shared screenshots of the original article on Imgur, allowing the public to see the controversial content that was published.

The article was credited simply to “Microsoft Travel,” implying a lack of human involvement in its creation. According to the “About Us” page on Microsoft Start, the platform claims to utilize “human oversight” in the algorithms that sort through a vast array of content provided by partners. This process is designed to assist the company in understanding factors like freshness, category, topic type, opinion content, and potential popularity, allowing them to publish according to user preferences. It’s worth noting that Microsoft laid off approximately 50 reporters from this division in 2020 while transitioning to a model relying heavily on AI-generated content.

Microsoft is not the first organization to experience backlash due to its reliance on AI-generated content. Earlier this year, CNET faced criticism for publishing numerous error-laden financial explainer articles created by artificial intelligence. More recently, Gizmodo’s parent company, G/O Media, published an AI-written article about Star Wars that was filled with mistakes, which deputy editor James Whitbrook described as “embarrassing, unpublishable, disrespectful.” While the Associated Press continues to approach AI-assisted news coverage with caution, other media outlets—including Microsoft’s publishing arm—appear to be more willing to embrace fully AI-generated articles, often addressing the fallout only after the fact.