News Summary

• Generative AI (GenAI) systems are increasingly recognized as cultural technologies.
• Current evaluation frameworks often treat culture as a variable to be measured rather than fundamental to the system's operation.
• We argue that GenAI systems function as "context machines" that must inherently address three interpretive challenges: situatedness, plurality, and ambiguity.
• We present computational hermeneutics as an emerging framework offering an interpretive account of what Gen AI systems do, and how they might do it better.
• This perspective offers a nascent paradigm for designing and evaluating contemporary AI systems.