From publishing to AI

Victoria McClune, a creative director, has tracked digital technology from interactive iPad books to artificial intelligence.

Noticing the system changed

McClune observed an AI model treating her differently after she introduced language about children and healthcare. Drawing on her psychology background, she tested the system to investigate what was influencing its responses.

“I noticed I was being managed and I knew what that was from my background.”

Silent Downgrade research

McClune's investigation became Silent Downgrade, examining how AI systems shape the quality of responses users receive. She argues that technical processes can influence experience before a user submits a prompt, in ways not immediately visible.

Users blame themselves

McClune uses a room metaphor to explain how AI establishes conditions influencing user experience. People assume disappointing answers reflect their inability to prompt effectively, without considering whether the system itself contributes to the outcome.

Industry neglects user experience

McClune criticizes frontier laboratories for concentrating on technical capabilities while giving insufficient attention to user experience. Businesses adopting AI need consistent outputs, yet confusing interactions make employees work harder to obtain reliable results.

More expertise needed in development

McClune wants professionals with expertise in psychology, communications and human-computer interaction to play a larger role in AI development. She is skeptical of relying solely on companies to regulate themselves.

Making systems more readable

At Current Futures Group, McClune focuses on making decisions and system states more readable to affected users. She hopes to help users understand AI without assuming they are at fault when interactions go wrong.

“I think what's actually missing is that there needs to be more people who understand us in the room.”