OpenBuild has been working along side Buildly and GetMotivatedBuddies for the last few months on an AI implementation for both platforms in a shared open source repo. One of the big headaches has been around #aitesting and #aiimplementation management in the long term. We decided to open source our own test plans and example scripts as way to share between us, but also with anyone else interested. We'd love to see others contribute what testing platforms they are using, how they are implemented and importantly, how they test post implementation that the responses still make sense and are ethical and fair to all users. Once the user is more involved in the training and weighting of the data through their own submissions, testing becomes even more critical and interesting and often more expensive.
We are starting signups for anyone who wants to contribute to the EthicalAI framework for testing and certification. It's a small open source project for now, but I think it will help startup platforms like ours stay ahead while implementing and testing AI. Please feel free to check it out, signup and include your GitHub username link if you want to contribute to the test plans, test script or just follow along. https://www.open.build/ethical/
We know we aren't the first to run into these problems or devise a solution but we feel strongly that we are going to the place to share those struggles and solutions. Together we will come up with a common way to make AI engines and AI implementations equitable and accessible to everyone while helping startups keep the costs of implementation down.