Fake reviews on AliExpress are real, but they are less common than useless ones — reviews written before the parcel arrived, reviews of a different variant, five stars with no text because the app asked. The skill is not detecting fraud; it is finding the twenty reviews out of five hundred that actually describe the product. This guide shows how to do that by hand in a couple of minutes, and how AliScore summarises the same review data into a positive-versus-negative split on the page, so you start from the numbers instead of from page one.
First, the honest framing
No extension, including ours, can tell you that a specific review is fake. Anyone who claims an "AI fake review detector" for AliExpress is guessing at scale. What can be measured is the shape of the review history: how many reviews exist, what share is negative, whether the structured Quality and Value labels buyers chose are positive, and how many reviews arrived with photos. Those numbers are hard to fake cheaply, and they are what AliScore reads. The judgment about individual reviews stays with you, and it goes faster once the numbers have told you where to look.
Six signs a review is not worth reading
1. It was written before delivery was possible
AliExpress shows the review date and, for many orders, the shipping route. A review posted two days after a listing went live, on an item shipped from overseas, was written by someone who had not received it. These are usually seller-side or incentivised, and they are always worthless.
2. It says nothing specific
"Good product, fast shipping, recommend." That could be pasted under any listing on the site, and often is. Real reviews mention a detail the reviewer could only know from using the item: the plug adaptor that was included, the strap that is shorter than pictured, the battery that lasted a week.
3. Five stars, no text, no photo
The AliExpress app prompts buyers to rate, and many tap five stars to make the prompt go away. These inflate the average without carrying information. A listing whose reviews are mostly star-only is a listing whose 4.9 tells you very little.
4. A burst of reviews in one week, all similar
Sort by date. A cluster of near-identical reviews posted within a few days — same length, same praise, same missing photos — is a pattern that organic buying does not produce.
5. The review describes a different product
Sellers reuse listings that carry a good review history. If the reviews praise the sound quality and you are looking at a charging cable, the reviews belonged to a previous product on the same listing.
6. Photos that are the listing's own photos
Buyer photos are the most trustworthy element of a review. But a "buyer photo" that is a crisp studio shot identical to the listing gallery was uploaded from the listing, not from a kitchen table.
How to find the reviews that matter
- Filter to reviews with photos first. A blurry photo of the item on a desk is worth more than ten paragraphs of praise. Look for the item in use, and compare it to the listing gallery.
- Read the lowest ratings next. One-star reviews are where the failure modes are: broken on arrival, wrong variant, stopped working after a month. Look for a repeated complaint, not a single angry one.
- Then read the three-star reviews. These are the most honest reviews on the platform — written by people who received the item, used it, and had a mixed view. They tell you what the product is actually like.
- Check your variant. Many reviews show which variant was bought. A five-star for the large size does not describe the small one.
- Look at the date range. A product with reviews spread across two years has been sold, used and reviewed under real conditions. One with all its reviews from the last month has not.
What an AliExpress review checker can honestly measure
The individual review tells you about one buyer. The split tells you about all of them. AliExpress publishes review statistics for each product — the positive-versus-negative share, the five-star share, and the structured "Quality" and "Value" labels buyers pick when reviewing. AliScore asks that public endpoint about the product you opened and uses the answer in two ways:
- The Sentiment component of the 0–10 score, worth 25%, starts from the positive share and is penalised once negatives climb past 10% of reviews — the point where complaints stop looking like noise.
- A named warning when negatives cross that line: "14% of reviews are negative — above the 10% mark where buyer complaints stop looking like noise." The same threshold that moved the score is the one that fires the warning, so the two never disagree.
It also shows the review count, and holds the score down when there are fewer than ten reviews, because a positive split over six reviews is not a split; it is six opinions. The complete formula is published, so you can check whether you agree with the weights.
Putting it together
A listing you can trust the reviews on has at least a few hundred of them, spread across months, with a negative share under 10%, a healthy number of buyer photos that match the listing, and three-star reviews that describe the product you are actually buying. A listing whose reviews fail those tests is not necessarily fraudulent — it may simply be new — but its rating is not evidence, and the price should be judged as if the item were unrated.
The rating system guide explains why the average itself carries so little information on AliExpress, and the seller ratings guide covers the store-level feedback that sits behind every listing.
Spotted a mistake or something out of date? Mail anwarazim03@gmail.com with the URL — one person maintains this and reads every message.