AI and ML for Games
From detecting fraud and predicting player behavior, to automating speech recognition and playtesting, Artificial Intelligence (AI) and Machine Learning (ML) makes games smarter and development faster.
Build intelligent game experiences
Featured use cases
Fraud detection before any player impact
Train and run your own ML models that help detect in-game fraud before it can impact your game’s lifetime value.
Establish a fraud detection pipeline
Fraud detection using machine learning enables you to execute automated transaction processing on a dataset to detect potentially fraudulent activity and flag that activity for review. The following diagram presents architecture you can automatically deploy using our implementation guide and AWS CloudFormation template.
Use this as an out of-the-box solution or a reference implementation for detecting player fraud on credit card transactions for in-app purchases, premium subscriptions, or in-game marketplaces.
AWS also offers a fully managed service through Amazon Fraud Detector.
Automated game leveling
Personalize level balancing for players using reinforcement learning with Amazon SageMaker.
Level design is an important part of creating a good game. If your level is too hard, players will get frustrated and stop playing. If your level is too easy, players will get bored and stop playing. As you add new levels to a game, you want to ensure they’re fair in an objective, quantifiable way. And you want to automate as much of the process as possible, so you can ship quickly. Instead of spending hours manually validating, testing, and fine-tuning your level design, automate the process with reinforcement learning.
Real-time sentiment monitoring
Monitor the sentiment of a Twitch chat in real time with Amazon Comprehend.
Understanding the sentiment of your player base is key in responding to feedback and building a game players want to play. Let’s say you’re getting ready to launch a risky new feature that you plan to test in a subset of channels. You have a good metric to optimize with an A/B test, but it’s a lagging indicator and you need to be able to roll back quickly if players get frustrated.
To reduce the risk, you build an application that monitors the sentiment of Twitch chats in real time. Then, you write a chat client that measures the sentiment of every message with Amazon Comprehend. The average sentiment by minute is displayed in an Amazon CloudWatch dashboard, so you can rest easy when your new feature goes live.
The concepts covered in this workshop can be applied to other data source like text chat, discussion boards, and other social media channels.
Automated image analysis
Save hours of manual data labeling by using machine learning to automatically tag images in your asset library.
If an artist takes 40 seconds to open a file, look at the image, form an opinion, and write multiple tags to it, then store it in a database, it would take them over 55 working hours to process 5,000 files.
Let artists focus on designing an immersive world for players. Using Amazon Rekognition, AWS’s image recognition that utilizes machine learning, asset labeling becomes simpler and exponentially faster. In our test code, uploading at a speed of 200 megabits per second, it took a little less than 3 minutes to complete image labeling.
Virtual voice actors
Use the power of the cloud to add lifelike voices into your games without having to hire voice actors, or pre-build in-game dialog before hiring a voice actor to produce the final voice track.
How virtual voice actors work
Leveraging the deep learning technologies of Amazon Polly gives you a quick and frictionless way to generate lifelike speech in your games, with support for 24 different languages and 50 unique voices.
What is Amazon Polly?
Amazon Polly is a Text-to-Speech (TTS) service that uses advanced deep learning technologies to synthesize speech that sounds like a human voice.
Not sure where to start?
Detect fraud in games using machine learning
Learn how to get fraud detection using ML up and running, so you can train and run your own machine learning models that help detect in-game fraud.
Identifying suspicious activity with an ML powered abuse detection pipeline
Learn how to leverage AWS machine learning and artificial intelligence to build a dynamic and performant end-to-end abuse detection pipeline.
Featured Customer Story
Meet studios that innovate for their players with AWS
See related technical guides, webinars, white papers and much more.
Discover how to answer difficult questions about changes to your game and player behavior using knowledge graphs and Amazon Neptune.