Fake AI-Generated Reviews Rampant in App Stores
DoubleVerify reports that AI-generated fake reviews on mobile and smart TV app platforms will triple in 2024 compared to the previous year. Fake reviews boost the rankings of low-quality apps, misleading users and exposing them to risks such as excessive ads and device control. Consistent writing styles, unusual syntax, high ratings, and written reviews for each rating are all red flags of fake reviews.
The company's Fraud Lab reported a significant increase in apps with AI-powered fake reviews in 2024, tripling the number from the previous year.
These deceptive practices have serious consequences, as low-quality apps with excessive advertising are unfairly boosted in app rankings. This misrepresents them as high-quality apps, potentially leading unsuspecting users to download them and expose them to risks.
DoubleVerify warns that some of these fraudulent apps may attempt to take control of users' devices, bombarding them with ads even when the phone appears inactive, thereby benefiting fraudsters while draining device batteries.
While DoubleVerify did not name specific apps, its investigation revealed numerous apps with artificially crafted five-star ratings, most likely generated by AI. Although fake ratings are not new, using generative AI makes them more sophisticated and difficult to detect.
Interestingly, DoubleVerify discovered a pattern in which AI-generated fake reviews used repetitive phrases across multiple reviews, particularly in smart TV apps, with all of them rated four or five stars.
DoubleVerify identified fake reviews associated with CycloneBot, a platform that generates fake streaming traffic to boost ad revenue despite the lack of real viewers.
The prevalence of AI-generated fake reviews provides a cost-effective strategy for dishonest app developers, as purchasing fake reviews can cost as little as US$2.23, limiting the number of reviews a developer must purchase.
An analysis conducted in 2023 discovered that a quarter of the top apps on Google Play and 17% on the iOS App Store in popular categories had suspicious reviews. Games and health and fitness apps were identified as common culprits due to their potential for quick profit.
To identify apps with fake reviews, DoubleVerify outlined five key indicators:
- Consistent writing style across reviewers.
- Unusual formatting or syntax.
- High four or five-star ratings
- Written reviews accompanying every rating, even if not required.
- Reviewers are exclusive to a single platform.
However, detecting fake reviews necessitates meticulous scrutiny of multiple apps and reviews, which most consumers are unlikely to undertake. As generative AI advances, the authenticity of fake reviews is expected to improve, forcing app stores such as Apple and Google to improve measures to verify review legitimacy.
DoubleVerify reports a sharp rise in AI-generated fake reviews on mobile and smart TV app platforms in 2024, tripling from the previous year.
Fake reviews boost low-quality apps in rankings, misleading users and exposing them to risks like excessive ads and potential device control.
Key signs of fake reviews include consistent writing styles, unusual syntax, high ratings, and written reviews on every rating.
Source: FORBES