Fake reviews flood the web
Amazon, Google and co. can't keep up with deleting fake reviews. The fake factories are hiding abroad and courts can hardly get to them. AI could soon make it even harder to spot fake reviews.
People who want to travel, are looking for a good restaurant or simply want to buy something often trust the experiences of other people. Around two out of three customers rely on reviews on online portals or personal recommendations, according to the German Retail Association's consumer monitor.
If the product has five out of five stars next to it, we immediately have a good impression. Authentic experiences of other customers influence whether we buy or not.
However, consumer protection experts warn against blindly trusting online reviews. A third or more are manipulated. And not all fake reviews are found: "Google and Amazon have removed hundreds of millions of fake reviews from their websites in recent years. The number of unreported cases is likely to be many times higher. Consumers keep reporting fake or supposedly fake reviews to us that companies don't discover," says Sabrina Wagner, digital market observation officer at the Federation of German Consumer Organizations in the ntv podcast "Wieder was gelernt".
Fake reviews cost the economy billions
Fake reviews have actually been banned in the EU since May 2022. The EU directive states that platforms may only publish reviews from real customers who have actually purchased the product. They must disclose whether and how they check reviews.
However, according to an investigation by the consumer advice center, only a fraction of providers have done this so far. In June, it examined 30 platforms, including Amazon, Lidl, Obi and Zalando, and rated 27 as inadequate.
Fake stars can be found practically everywhere: on the large portals as well as on smaller online stores or medium-sized companies. But especially where a lot of money is involved. The rating service Trustami estimates that false ratings cost the German economy 3.8 billion euros every year.
Court condemns rating agency
There is a complex system behind fake reviews. Retailers usually commission review agencies. These in turn hire individual reviewers. They write what the agencies tell them to - sometimes without ever having had the products in their hands. "This can be worthwhile, also because the testers are paid little to miserably," says Wagner. "Ten positive fake reviews are worth 100 euros. The testers often get less than a dollar of that. That leaves quite a bit left over. In some cases, testers are remunerated with the products they rate."
Review agencies offer their services quite officially on the Internet. A single review on Trustpilot, for example, is available from the review company Fivestar for just under 15 euros, while a Google review costs around 13 euros.
One of the world's largest agencies for fake reviews on the internet is Goldstar Marketing. In October, the Munich Regional Court ordered them to cease and desist and pay damages. The travel booking portal Holidaycheck was able to prove to Goldstar Marketing that it had sold hotel reviews on its website without anyone having actually stayed there. If it does so again, it faces imprisonment or fines of up to 250,000 euros.
Company boss in hiding
Holidaycheck celebrated the ruling as groundbreaking in the fight against review fraud. However, it is of little use to consumers. This is because the judgment will probably never become legally binding. It would first have to be served.
Officially, Goldstar Marketing is based in Cyprus. However, the man behind the agency, Norbert Weber, has disappeared. He has not appeared in court. Holidaycheck has hired private detectives to find out his new address.
A well-known method used by dubious review agencies: when courts become aware of them, the company name is quickly changed or the boss replaced.
Fake review agencies also operate in a gray area, recruiting reviewers in special groups on messenger services such as Whatsapp, Facebook or Telegram, reports consumer protection expert Wagner in the podcast. "This makes it much more difficult to get to the people behind it." Amazon regularly tries to take action against this, has the relevant groups closed down and sues the operators. From January to May, the company said it took legal action against dozens of providers of fake reviews in the USA, China and Europe.
Alliance against fake reviews
Amazon is barely able to combat fake reviews. Last year alone, the world's largest online retailer deleted 200 million suspicious reviews. To get a better handle on the problem, Amazon joined forces with industry giants Booking.com, Expedia, Glassdoor, Tripadvisor and Trustpilot in October. They want to stop fake reviews. To this end, representatives of the platforms will meet at their second fake reviews conference at the beginning of December.
However, Wagner suspects another background: in the USA, the FTC is planning a ban on fake reviews. Penalties of up to 50,000 US dollars are threatened for manipulated reviews. Skeptics doubt that these measures will have any effect. This is also because many of the review companies - as in the case of Goldstar Marketing - are based outside the USA. This makes prosecution difficult.
How do I recognize fake reviews?
It is difficult to recognize fake reviews - but there are a few clues. In the "Wieder was gelernt" podcast, Wagner warns against being impressed by too many positive reviews. It can also be a conspicuous sign if language patterns are repeated in the text. "For example, if a robot vacuum is described as having 'excellent suction power' and this pattern is repeated again and again. The author's profile is also a clue: if they have written reviews for a wide variety of companies in one day, this could be an indication of contract work.
An anonymous profile or a pseudonym can also be an indication of a fake review. And if the reviews are written in broken German.
The portals use automated checking processes to filter out false reviews. With artificial intelligence, however, the software could reach its limits. Not only because of the sheer volume of reviews that are quickly created with ChatGPT and the like. With the help of AI, fraudsters could also write their fake reviews in such a way that they can no longer be distinguished from genuine reviews.
Source: www.ntv.de
