Spirit AI wants to be your Ally on the fly, a player-centric bot for online gaming abuse

Spirit AI wants to be your Ally on the fly, a player-centric bot for online gaming abuse

Online harassment or abuse in gaming involves some of the most toxic slurs ever slung in cyberspace. For more sensitive gamers, simply saying, “Don’t take it personally,” doesn’t cut it.

No race, creed, ethnicity, gender, nor religious affiliation is exempt from verbal or textual abuse online. Although many gamers see this as just virtual trash-talking and part of the territory, a poll from IGN shows that one in three gamers are actually turned off by online abuse.

Read More: Google’s new Perspective API can help you not sound like a jerk while commenting

“While the use of boastful or insulting speech to intimidate or humiliate can have value as a psychological strategy, when the remarks attack someone for their gender, perceived sexual orientation, or race, many would agree that a line has been crossed,” wrote Kaitlyn Williams, who received a Stanford University Boothe Prize Honorable Mention for her essay When Gaming Goes Bad: An Exploration of Videogame Harassment Towards Female Gamers.

In order to evaluate whether a line has been crossed in online gaming interactions, the team at Spirit AI developed a bot called Ally that makes “in-game player communities, chatrooms and online social platforms safer and more inclusive.”

Using the power of machine learning and predictive analytics, Ally takes a player-centric approach to abuse, asking the player whether or not they are OK with player interactions and learning what situations, language, and individuals are within their “safe zones” or comfort levels.

What’s cool about the player-centric approach is that Spirit AI’s Ally comes in the form of a virtual character that mimic’s the game’s style, so it doesn’t seem out of place.

The Ally checks-in on a potential abuse victim and asks whether or not a user has been offensive, and the software has a customizable interface that allows players to input what is offensive to them.

On the surface this may seem tedious and even overkill on political correctness.

However, the Spirit Triage Manager will decide how the system will respond to each abusive scenario as it is detected, and its context-aware reporting system can create a case file for further analysis by the community team, whether a player proactively reports an instance of abuse or responds to an Ally enquiry.

It’s good to know that there is at least some human element working behind the scenes.

Additionally, as players gain more experience in a game or chat room, or build their own community with whom they feel relaxed, their response to problematic language may change. They’re free to tell Ally whenever their preferences change – or even how they’re feeling on any given day.

Click to add a comment

Leave a Reply

Your email address will not be published. Required fields are marked *


Tim Hinchliffe is a veteran journalist whose passions include writing about how technology impacts society and Artificial Intelligence. He prefers writing in-depth, interesting features that people actually want to read. Previously, he worked as a reporter for the Ghanaian Chronicle in West Africa, and Colombia Reports in South America.

More in Technology

ai citizen happiness

AI for governments monitoring citizen happiness

James RosenfieldMarch 21, 2018
ai task force china

AI Task Force Rhetoric Encourages Cold War With China Over National Security

Tim HinchliffeMarch 19, 2018
cryptocurrencies storm

Cryptocurrencies: A Perfect Storm

Melissa ThompsonMarch 16, 2018
native advertising

The rise of native advertising, and what the future holds: Interview

Sam Brake GuiaMarch 16, 2018
artificial intelligence startups

The Rise of Artificial Intelligence Startups

John MasonMarch 16, 2018
ai defense

The AI revolution in defense technology is happening faster than we think

James RosenfieldMarch 14, 2018
mossad invest startups

Why does Israeli spy agency Mossad invest in startups via Libertad Ventures?

Tim HinchliffeMarch 12, 2018
emotional intelligence

Artificial vs Emotional Intelligence in Machine Learning

James RosenfieldMarch 12, 2018

Cyberbullying isn’t going anywhere; here’s how to deal with it

Melissa ThompsonMarch 12, 2018