Stepford Project Timeline

Naromass receives funding from Mozilla Technology Fund. Announcing the winners of The MTF: Bias and Transparency in AI Awards, Mehan Jayasuriya, Senior Program Officer at Mozilla says, “All five of these projects are breaking new ground in the AI transparency space”.

The first phase of project development looked in detail at the requirements for development, front-end design and began development of backend infrastructure, such as back-end tooling for text corpus conversion and API communication management, user authentication integration. We also developed a corpus of test fiction text segments to help standardise testing.

At Mozfest, we presented the background of the project, including what we learned from making the first Algowritten short story collection. We showed a video of the working Stepford prototype that took an AI-generated text segment and used GPT-3 to analyse it for sexism and declare a reason for its assessment.

Video of the app in development, used at MozFest 2022

The second phase of project development focused on making the Stepford application, easy to use in terms of front-end design and secure and robust at the back-end. In this phase of the project we also tested and developed the agent scripts that were used to make Stepford able to select sexist comments and have an opinion on them. Internal and then friends and family testing was used to test the app before launching testing officially at the event in November.

In September 2022, we were invited to present Stepford at the Algorithmic Reparation Workshop at University of Michigan by Apryl Williams, a Mozilla Fellow we met during one of our monthly monitoring meetings. We were excited by the opportunity to learn about Algorithmic Reparation and to receive feedback on Stepford’s development, so we jumped at the chance.

In November of 2022, Stepford was presented to and tested by a group of eight volunteers in Manchester, UK.

Naromass gave a short presentation on the development of the tool, including theoretical influences and the representation of AI in media.

In self-selected pairs, participants engaged with the tool, indicating whether they agreed with Stepford’s scoring of texts for sexist bias.

The Stepford demo will return virtually to the University of Michigan in the spring of 2023 as part of a class on Critical Internet taught by Apryl Williams.

We are eager to get this tool into the hands of young creators at the School of Digital Arts in Manchester, UK, where we plan to host an event this year that will encourage both creative experimentation and critical feedback.

One of the features we plan to add to Stepford is a free-text box, where users can write or paste in their own text for analysis by Stepford. We anticipate this feature will widen its usability and provide a greater breadth of training data to improve the tool.