Research
Evidence from real creative work
We work directly with AI systems through practical creative tasks, controlled comparisons and sustained collaboration. Our analysis considers not only what a model produces, but how it interprets the brief, responds to feedback, preserves human intent and affects the work required from the person using it.
Current study
One Brief, Seven AI Models
Seven leading AI systems were given the same live editorial brief about DLSS. The study compares their research, accuracy, reasoning, structure, voice and ability to transform the original work under new tonal direction.
Status: Complete. Report 001 compares 21 responses across the initial, humorous and confrontational rounds.
Read Research Report 001 →Publications
Latest research
One Brief, Seven AI Models
Seven AI products answered the same live DLSS brief, then rewrote their own essays with humour and confrontation. Report 001 compares research, adaptation and editorial control across all 21 responses.
Method
Transparent comparison
Prompts, outputs, model versions, tools, interventions and limitations are made visible wherever practical. We distinguish technical capability from creative judgement and record the human work required to turn generated material into something accurate, distinctive and useful.
- Creative capability
- Model behaviour
- Human–AI collaboration
- Creative tool tests