Riot staff was caught using AI generated response after a reddit user’s support ticket was returned with the phrase ‘we at Character.AI …’. The response was later edited to ‘we at RIOT …’, and a different Rioter had responded to clarify the hilarious situation.
The age of Artificial Intelligence is upon us and it’s of no surprise that a smartass Riot staff would use it to their advantage.
Character.AI is an AI chatbot website where users could modify it to act like any character they define. This ranges from fictional characters to real life celebrity, and in this case a Customer Support worker.
There have been increasingly cases where students and workers alike are caught using chatbots to do their homewor/jobs for them. Leaving aside regulations, work and moral ethics aside, this does result in some fantastic hilarity.
Riot Castmoore clarified that, no information from players have been passed to Character.AI or any 3rd-party chatbots and reassured that Riot staff is not using AI to generate responses to support tickets:
I’m Riot Castmoore and I oversee some of our work on Player Support. ChatGPT and similar AI emergent tech is evolving rapidly and we haven’t had this issue come up before so I wanted to look into it personally.
We’re still looking into exactly what happened here, but the short answer is no. Individual agents often use personalized macros to help respond to players more quickly, but we do not write responses using AI, or submit any player information into any AI platform. When we do use AI in the player support process, we’re up front about it. It looks like, as part of prepping their ‘macro’ response for questions like this, this agent used some AI-generated language. This is not part of our best practices or protocol, and should not have happened.
You may have seen some automated responses from Blitzcrank Bot and our online chat widget which identifies and discloses that it is AI. This is about 15% of all ticket responses.
We do not use Character.AI or have any relationship with them, nor do we use other AI’s like ChatGPT to generate responses to players. And just to be super clear, we don’t allow player requests to go through any tool outside of our secure network. This was an individual agent error and shouldn’t have happened. That being said, we’re all adapting to the emergent AI landscape and there’s a lot of takeaways and learnings from this interaction.
Thank you for reaching out about this.
It seems that in this particular instance, the staffer merely used the chatbot to generate a prompt, which they then use to set up an automated macro for support tickets. The macro is common practice at Riot, but using a chatbot to create the macro is frowned upon. Riot is still looking into this incident as of this moment.
The reddit has since generated a lot of tractions with over 8.5k upvotes as of writing.
Speaking of support, buying support items somehow is currently meta for top laners