Artificial Intelligence, or AI, has been very beneficial in many areas, but it can also be used in less-than-salubrious ways. In 2023 alone, it was reported that Google took down an eye-popping 170 million fake reviews¹ of hotels, restaurants, and businesses, which was a 32% increase from 2022, while Trustpilot flagged 3.3 million reviews² as fake on its site.
With the wide-scale adoption and access to AI technology, fake reviews are becoming more prevalent and sophisticated, and they are becoming tricky to spot. These fake reviews are blurring the line between authentic feedback and fabricated content.
With summer sales in full swing, AIPRM’s AI experts have put together a guide that will help you spot fake AI reviews and should provide you with more confidence while shopping this summer.
Watch for Generic Gimmicks
AI-generated reviews can sometimes seem like they were pulled from a generic, one-size-fits-all script – bland, vague and lacking that human spark. Be wary of reviews that rely on generic phrases and could be applied to almost any product.
Instead, look for reviews that describe how the product was used, highlight specific features that stood out, and share any unique experiences the reviewer had. Genuine reviews often include nuances and personal insights that are harder for AI to replicate.
Example:
Potentially fake review: “This product is great. It works well. I recommend it.”
Real review: “I’ve been using this vacuum cleaner for two months, and I’m really impressed with how it tackles pet hair! The battery life could be better, but overall, it’s a solid purchase for pet owners.”
Analyse the Reviewer’s Profile
Get into detective mode and scrutinise the reviewer’s profile for clues about authenticity. Real users usually have a history of varied reviews across different products and services.
In contrast, fake profiles might only have a handful of reviews, often clustered around the same dates, or lack any personal details like a profile picture and bio.
Example:
Potentially fake profile: A user with three reviews that all have 5-star ratings posted within the same week.
Real profile: A user with a mixture of positive and critical reviews spanning various product categories over several months or years.
Look for Extreme Opinions
Another tell-tale sign of fake AI reviews is overly positive or excessively negative feedback that lacks any real details to back it up. Genuine reviews often present a balanced perspective, acknowledging both the pros and cons of a product or service.
Example:
Potentially fake review: “Best product ever! No complaints at all!”
Real review: “I love the design and functionality of this blender, but it is quite noisy. It gets the job done efficiently, though.”
Cross-reference Reviews Across Multiple Platforms
Don’t just rely on one website for your review research. Be savvy and compare reviews across different platforms like Amazon, Yelp, and Trustpilot to get the complete picture.
Consistent feedback across different sites can indicate genuine experiences shared by real-life users.
Example:
Potentially fake review: A product with exclusively 5-star reviews on one site but a mixed range of ratings on others.
Real review: A product with a general trend of positive feedback and some critical reviews highlighting minor issues.
Examine the Timing of Reviews
If you notice a flood of reviews for a product all posted within a short time frame, especially if they are uniformly overly positive or negative, it might be a sign that AI is at play.
Fake AI reviews often appear in clusters, unlike authentic reviews that trickle in over time as users share their unique experiences at different intervals.
Example:
Potentially fake review: A product receives dozens of glowing reviews, all posted within a few days or hours.
Real review: A product has reviews spread over several months, with users sharing their thoughts at different times.
References
¹ https://timesofindia.indiatimes.com/gadgets-news/google-removes-170-million-fake-reviews-in-2023-with-new-machine-learning-algorithm/articleshow/107698607.cms
² https://www.techdigest.tv/2024/05/trustpilot-removes-3-3-million-fake-reviews-in-2023.html
You must be logged in to post a comment.