Opinion] Project Ava, Razer’s new gaming AI, is causing Jusuf to worry about his favorite multiplayer game. Is this really the future of League of Legends?
CES 2025 may already be two weeks behind us, but I still haven’t digested one of the announcements made at the technology trade show.
It’s about “Project Ava,” Razer’s new AI tool that positions itself as an AI-powered gaming assistant. With this, the company wants to give you analyses of your completed games, optimize your hardware, and yes, even support you during the game.
AI assistance in single-player is not the problem
Whether you use Ava in single-player adventures like the one shown in Black Myth: Wukong is not important to me. In my opinion, everyone here can get the help they need. I don’t care whether you want to finally defeat that wretched level boss or whether you just want to have more fun playing the game.
That’s why we play (most of the time): to have fun. In this respect, the Razer AI is a useful tool and serves a similar purpose to our special editions whenever a big game comes out.
With the difference that our guides to Elden Ring, Civilization 7 & Co. are of course more qualitative and ingenious.
And what happens in single player stays in single player; let him who did not enter cheat codes like “rosebud” in the first Sims or “iddqd” in the first Doom throw the first stone at those who did so back in the day.
My problem, therefore, lies instead in a short section that Razer showed in the teaser: Apparently, Ava can also be used in multiplayer games like my beloved League of Legends. The developers call this “real-time coaching”. Depending on the situation, Ava is supposed to give you hints on tactical decisions during the game.
Imagine Ava calling out to you at a critical moment in the first few minutes of League of Legends: “Careful, the enemy wants to overpower your lane – retreat and prepare for evasive maneuvers.”
Or how Ava warns you in Call of Duty about the sniper who is hiding half a kilometer away. “An enemy scope flashed in the ruins to your left, run quickly!”
Is this really the future that awaits us in our multiplayer games? I hope not.
Because that’s exactly what makes such competitions so appealing to me: competing against other players. How do my strategies work? Are my decisions right or wrong on a tactical level? Are my reaction times faster and more precise than those of my opponents? (Answer: No. I’m not getting any younger either.)
“But Jusuf,” I hear you call out, “didn’t you just last year celebrate an AI-based tool specifically for League of Legends?”
Secondly: Yes and no. The software “iTero” I presented back then provides analysis before and after a game, but never during the game.
This is where I see the difference. Whether I chew over patch notes or go through the numerous overview pages for victory percentages and suggested items – as long as I do it as a kind of “training” away from a live game, I see no problem with that.
But since Project Ava, at least according to the Razer announcement, wants to intervene in the game in progress and issue warnings, that’s too much of a good thing.
Not every League assistant is a cheat
But to be on the safe side and get another perspective, which is on the other side of the AI discussion in games, I grabbed Jack Williams and engaged him in a short conversation.
- Williams is the founder of the aforementioned iTero and “Head of Gaming Tech” at the professional LoL team GiantX, which iTero acquired in a multi-million dollar deal last year
- iTero also has an analysis tool that uses machine learning to optimally prepare data such as win rates or correct runes before the game. Afterwards, there is another analysis of your profile to point out your statistical weaknesses.
According to Williams, the crux of the matter lies in the following aspect: Where does the data come from that tools like iTero, Mobalytics, Blitz and all the others use as a basis – and where does it come from from Project Ava?
The first group receives the data sets from the official developer interface of League maker Riot Games. This also allows the developers to control exactly which data is and is not allowed for such an analysis before and after the game – so that tools like iTero, Blitz or Mobalytics also receive their “Approved by Riot” seal.
Williams does not dispute that the use of such applications in competitive multiplayer is also considered a competitive advantage despite Riot’s apparent approval – but since the publisher itself promotes such projects, the cheat accusation is rather untenable.
- In contrast, Project Ava (at least according to the announcement) partially analyzes the running screen content.
- This means that data that is not provided by the developer in either form or quantity is analyzed at a speed that the human player cannot keep up with.
Almost the same discussion again, almost to the day
At least in view of this argument, it makes sense to me, especially in the “league bubble”, why software like iTero, Blitz and others are regularly advertised and used. In combination with the fact that these tools do not interfere with the game in progress, Project Ava is already under observation by me.
The whole thing is somewhat reminiscent of the MSI monitor that was presented at CES 2024 and also triggered a cheat discussion. To put it simply, “AI Skylight” was designed to turn warning lights on your display when opponents sneak up on you from the side. To do this, the minimap was also analyzed.
The end of the story: MSI canceled the tool for multiplayer games, and the project is now only in development for singleplayers. However, we won’t know whether this scenario will also happen to Project Ava until March at the earliest, when Razer starts the beta phase for AI assistance.
Riot Games itself has not commented on this either. I hope that this will change soon and that I will have certainty about what not only we League players, but also the entire multiplayer scene, will have to adjust to.
