Promoting Equality through Appian

Appian based case management solution that uses machine learning to detect unconscious bias to help improve your organization's decision making.

The inspiration for this project was the idea of using technology to fight against bias in society. There have been a lot of articles about how automation and AI can perpetuate bias by mimicking the biased patterns in their human generated training data. We wanted to reverse that trend, and use technology to eliminate the bias in human generated data instead of perpetuating it. To do this, we combined Appian and Rhino AI to create a collection of business processes which would automatically be audited for bias.

Appian's ability to automate the processes of almost any industry make it an excellent platform to build this functionality in. If bias can be reduced through Appian processes, it can be reduced in all the business and enterprise areas that Appian processes are used. This has the potential to create enormous societal impact.

Some challenges we had to overcome were how to identify different types of bias in unstructured data, such as racial, national, or sexual bias. To do this, we used AI models hosted in Rhino, which was trained on biased text from both human sources and machine sources such as GPT-3. Due to Appian's excellent integration capabilities, it was easy for us to integrate Rhino's AI bots into Appian's processes, allowing us to rapidly add auditing capabilities to multiple processes. We could then display the AI bots‚Äô findings through Appian's rich UI and reports.

We look forward to continuing this work in the future, and believe that this can be used as a template to help reduce bias both in the enterprise and in society at large.