Microsoft’s news platform MSN infamously fired a large portion of its journalists in 2020, replacing them with artificial intelligence. Two years on, those AI journalists are creating harmful misinformation that no one bothers to check.
AI text generators have evolved substantially in the last few years. However, they still have issues with disinformation. For example, a recent Meta AI created full science papers riddled with false facts that were wrongly attributed to real-life scientists.
MSN may be even worse, though. While not as inherently dangerous as a science paper, there are more eyes on Microsoft’s “news” platform than your typical academic study.
In a report by Futurism, MSN’s actual fake news is a huge part of the AI-generated outlet. Recently, the outlet published a fake story about musician Grimes calling out ex-husband Elon Musk for not paying child support, but that’s the tip of a much deeper iceberg.
As it turns out, MSN’s source-crawling picks up a number of problematic stories it then regurgitates as news. For example, the outlet often pulls from a website called Exemplore, an outlet that “reports” stories regarding popular conspiracy theories such as cryptids.
Exemplore articles frequently discuss UFOs, Bigfoot sightings, Angel appearances and much more. MSN’s AI journalists see this as actual news instead of the obvious bogus it is. This means that the website is filled with bogus articles spreading false stories.
MSN is not a small-scale website. While it has fallen into obscurity for many, the website is still used every single day. Importantly, the outlet is embedded into Windows — opening the default Windows browser shows the homepage for MSN. Even using the web browser on an Xbox console will open the website.
Furthermore, the website is a legacy website which means it’s more likely to have its stories pulled into Google News and Google Discover. With decades of domain authority, the website is still primed to take up Top Stories boxes on search engines and deliver misinformation to the masses.
There’s a reason why actual journalism is performed by human beings. At its current level, AI cannot properly determine fact from fiction. As such, it sees these fake stories as notable events that must be spread online, and that’s inherently dangerous.
With the massive output that Microsoft’s news platform has, it’s incredibly dangerous to have AI roam free. Furthermore, as one of the biggest companies in the world, Microsoft should have the funds to pay human writers to create actual, fact-based content.