Closing the Developer AI Trust Gap

Mind the gap: Closing the AI trust gap for developers

The Stack Overflow 2025 developer survey reveals a significant drop in trust in AI tools among developers, with only 29% of respondents saying they trust AI. This is a sharp decline from 2024, when 40% of developers reported trusting AI tools.

Why This Matters

The trust gap between usage and trust of AI tools among developers matters because it impacts the productivity, scalability, and innovation potential of AI. If developers do not trust AI tools, they will not deploy AI-generated code to production systems, which can lead to missed opportunities for growth and efficiency. Furthermore, the lack of trust can also lead to increased technical debt and risks, as developers may be more likely to rewrite AI-generated code or avoid using AI tools altogether. For instance, if a critical system fails due to an AI-generated code error, the cost of repair and recovery can be substantial, highlighting the need for developers to trust AI tools.

Key Insights

• Stack Overflow’s 2025 survey found that 84% of developers use or plan to use AI tools, but only 29% trust them
• Developers are trained for deterministic thinking, which can lead to mistrust of probabilistic AI tools
• Effective prompting and evaluation frameworks are crucial for building trust in AI-generated code, as seen in Uber’s approach with Genie

Practical Applications

• Use case: Uber’s Genie AI assistant, which provides accurate and reliable answers to questions in Slack channels and resolves support tickets, demonstrating the potential for AI tools to enhance productivity and efficiency. Pitfall: Failing to implement proper evaluation frameworks and testing for AI-generated code can lead to errors and security vulnerabilities.
• Use case: Stack Overflow’s own AI-powered coding tools, which can help developers with tasks such as code completion and debugging, highlighting the potential for AI to improve developer productivity. Pitfall: Over-reliance on AI tools without proper understanding of their limitations and potential biases can lead to decreased code quality and increased technical debt.

References:

• https://stackoverflow.blog/2026/02/18/closing-the-developer-ai-trust-gap/