AI defects
Explore failure modes, the checks TextData demonstrates and the limits of their evidence.
Powered by frontier AI
From text to rigorous, novel organizational research.
Bring sources, methods and tools together—and adapt them to your research question.
Reversed movable type.
Honneur à la théorie.
Salut aux savants.
Existing components and worked examples are available to explore. Further integration and evaluation are planned.
Find your starting point
Choose the task you are working on to explore relevant materials, tools and examples.
Find sources and methods
Explore methodological literature, AI research systems and textual-data sources to judge what fits your question.
Records distinguish reported capabilities from proposed uses. Access, coverage and suitability need assessment for each study.
Develop qualitative explanations
Inspect source-linked coding, comparisons and revision records for developing explanations and examining contrary evidence.
Interview analysis is documented. Gioia examples use constructed material, and case analysis remains proposed.
Acquire and annotate text
Examine acquisition workflows and passage-level annotation through their saved inputs, source checks and inspectable outputs.
Acquisition demonstrations cover selected sources. Annotation replays saved responses; live model integration and independent reference coding remain proposed.
Design textual measures
Explore methods and measurement specifications, define what a construct means in text, and inspect how candidate annotations become features.
Specifications and software checks are available. Independent validation and a complete measurement workflow remain proposed.
Develop the capabilities your inquiry needs
The deep-research engine investigates methods, AI capabilities and technical repositories to inform what TextData should adopt, adapt or build. Explore a worked inquiry and the evidence it retained.
The worked inquiry demonstrates part of the design. Wider integration and comparisons after model changes remain proposed.
Inspect, adapt and contribute
TextData is designed to grow through shared tools, workflows and research materials. Explore the repository, inspect the available work and consider how its components could support your own inquiry.
Versioned open-source releases and a community contribution process are planned. The repository records what is currently available.
Examine the evidence. Know the limits.
Follow a design response to its supporting example and the questions it leaves open.
When AI goes wrong in textual research
Read how errors can enter an inquiry, how corrections can change its instruments and which workflow responses still need testing.
