Google sees you when you’re trolling
The future of moderating toxic online activity is in the hands (or teeth, depending on your perspective) of machine learning. If the term’s meaning is fuzzy — it’s basically artificial intelligence (AI) allowing computers to learn about their environment by exploring patterns in data. In the hands of search behemoth Google, machine learning got a real workout in 2021 by policing abusive behavior on Google Maps.
In a recent post about how Google keeps Maps information reliable, the company said it used machine learning and human operators to block more than 100 million attempted fraudulent edits to Google Business Profiles. It also took down more than 7 million fake profiles — and 630,000 of those removals were based on user reports. Additionally, Google says it also stopped 12 million attempts to impersonate a company and shut down 8 million fraudulent attempts to claim a company’s profile. On top of all that, Google shut down 1 million accounts due to related policy violations.
The report on Maps reliability also included measures Google took as businesses began to open back up in 2021 and people attempted to harm them with fake reviews. Over 95 million reviews violating Google policies were dumped, and 60,000 of those were related in some way to issues surrounding COVID-19. Machine learning also assisted the Google Maps team in taking down almost 200 million poor-quality or policy-violating photos and videos.
Google Maps can be a magnet for shady behavior, so it’s good to see how hard Google is working to clean house. But while Google Maps is using machine learning to great effect, it’s also good to know that human eyes are still involved in the process — especially if you feel that review joking about blowing up a pizza joint for its terrible garlic bread was unjustly deleted. Artificial intelligence has no sense of humor.
Read Next
About The Author
