AI-generated fake reviews mislead consumers and harm businesses
Fake online reviews created with AI assistance have become harder to detect and more effective at manipulating consumer purchasing decisions, costing consumers an estimated US$770 billion in unwanted purchases in 2025 alone while damaging honest businesses trying to compete.
What happened
According to March 2026 data from Capital One Shopping, fake reviews cost consumers worldwide US$0.12 per every dollar spent, amounting to approximately US$770 billion in unwanted purchases in 2025 alone, with high-stakes services in home repair, legal and medical fields making up the bulk of losses.
AI has made fake reviews significantly harder to detect and more effective. Review farms now generate natural-sounding copy and easily manipulate photos to make fake reviews look authentic, doing so at large volumes. This technology has tipped the scales in favour of businesses that use fake reviews, with the Capital One report finding that fake reviews can grow product sales by 12.5 per cent in the first two weeks alone.
Reported impact
Not only do consumers get duped by purchasing products or services based on fake reviews, but the practice also harms honest businesses trying to compete in environments where fake reviews are prevalent.
Index analysis
According to a consumer advocate and former U.S. federal criminal investigator, vast networks of fake reviews have been identified, many of which are based out of South Asia, but there are also plenty in Canada and the U.S. These marketplaces openly advertise services to buy, sell and trade reviews in bulk through Facebook groups and other social media.
Sophisticated marketing agencies also offer tailored approaches, with one North American company paying individuals $4 for a single Google review. In some cases, even 'elite' reviewers on certain platforms trusted for their consistent posts and accomplishment badges were approached to post fake reviews for cash.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.