Valve is updating its User Review system to prioritize relevant reviews that "best help players make a purchase decision about the game." Valve calls it the 'Helpfulness System,' designed to push one-word reviews, reviews full of ASCII art, memes, and unrelated information to the bottom of the list.
Valve won't remove reviews deemed unhelpful, and the new helpfulness system can be toggled on or off within the user review section of each game - by default, it will be enabled.
"The primary goal of Steam User Reviews is to help potential players make informed decisions about the games they are considering purchasing by understanding the attributes of the game that other players like or don't like," Valve writes. "Historically, we've sorted reviews by the number of 'helpful' votes given to each review by other players. However, we've seen that many players use reviews for sharing jokes, memes, ASCII art and other content that might not be the most helpful for a potential purchaser."
Valve adds that this change won't impact how it calculates game review scores, so the review bombing will continue for titles that release updates and patches that don't please the community.
So, how will Valve determine if a review is 'unhelpful'? It will use user reports, the Steam moderation team, and even "machine learning algorithms to help scale the human judgment calls." Yes, Valve will use AI to figure out which reviews are helpful to potential customers.
With so many games on Steam and so many reviews, Valve also says that it will take "quite a while" for them to go through and evaluate every single review. However, unhelpful reviews that are easy to spot are being targeted first.
