Model control, fine-tuning barriers, and AI system reliability

Research examining model control mechanisms and fine-tuning adoption barriers

<aside> ➡️ Contribute to the research

</aside>

<aside> 📝

Participants will be credited as research contributors.

</aside>

Research Question

Commercial AI solutions predominantly rely on a small number of foundation models (OpenAI GPT, Anthropic Claude, Google Gemini). This study wants to investigate the technical and organizational factors that prevent enterprises from developing more controllable, domain-specific AI systems.

Methodology

Semi-structured interviews with AI practitioners across three categories:

• Engineering teams implementing LLM-based applications
• Technical decision-makers evaluating AI solutions
• Organizations that have attempted model fine-tuning or training

Contribute to the research