Square Enix fans complain about machine learning tests

Square Enix recently released a free game on Steam called “Square Enix Test Preview: The Portopia Serial Murder Case”. This remake of his 1983 text-based adventure game used natural language processing (NLP) to correct what the publisher saw as limitations of the game’s original input-based gameplay. However, users are not happy with this new feature, to say the least.

NLP is a type of artificial intelligence that helps machines understand natural or spoken language. On the other hand, according to game reviews, Portopia isn’t a good demonstration of the technology simply because it doesn’t seem to understand natural language. This seems to indicate that even if the AI ​​and his NLP can be used in games, it’s still not quite where it should be, at least when it comes to how much agency its creators are willing to give players. .

What is the Portopia serial murder case?

The Portopia Murder Case is a murder mystery adventure and one of the most influential games in the visual novel and 2D adventure game genres. It was originally developed by Yuji Horii and published by Enix, the predecessor of modern Square Enix. In it, the player takes on the role of a detective trying to solve a murder case in his 1980s Japan with the help of his partner. This technical his preview is Portopia’s first release in English.

As described by Square Enix, the original version of the game relied on “command inputs”, which required the player to enter commands into the game. This page acknowledges the limitations of this. The problem was caused by the performance limitations of his PC at the time and NLP technology. ”

Portopia’s updated port does not eliminate command input entirely. Instead, use NLP to make it easier for the player to execute commands. According to the company, in-game natural language understanding (NLU) “helps junior detectives understand the player’s instructions.” In theory, this allows the user to get the desired response even if the user’s command doesn’t match exactly what the game expects.

NLP didn’t seem to solve the problem as much as the publishers hoped. At the time of this writing, the game is rated “very bad” on Steam. The general temperature from the reviews is that tech previews and natural language processing don’t seem to understand natural language. I had to simplify.

So what went wrong?

Reviews of the game on Steam show that the game frequently misunderstands or ignores commands, even though the commands are only slightly different from the intended commands. It looks like the AI ​​is parting the hair. If they don’t understand what you’re saying, your partner will respond in-game and tell the detective to focus on the task at hand.

One reason this seems to have failed is that Portopia doesn’t offer many text-based services. freedomPlayers must follow the story set by the game. Until the player hits the correct one, the inputs are always a matter of guessing and the game doesn’t progress until they hit them. As such, developers cannot give players a lot of agency regarding their own speech. Players “must say” certain things in order to advance the plot.

Another example of successful use of NLP in a different kind of game was in 2005 when a game called Facade used AI to tell a story in which the player was an active agent. They can say anything they want and the characters in the game will react accordingly.Even if the player doesn’t type a specific sentence, her two characters in the game will be able to communicate with her NLP in the game. I get the point thanks to This allows the player to freely explore their options and the story continues even if the game doesn’t fully understand the player.

At the very least, this seems to indicate that AI, or at least natural language processing, doesn’t always work in every game. Visual novels don’t give players much control. Unlike Facade, Portopia doesn’t allow players to say or do whatever they want. So his AI in the game can only respond to certain phrases. However, even within these limits, Portopia’s NLP looks inelegant and inelegant.