Beyond Chatbots: Bad AI Is Worse Than No AI

Adding AI to a solution doesn’t automatically make it better. In many cases, it makes things more confusing, less usable, and harder to maintain, especially when the underlying problem was never clearly defined.

In this session, we’ll take a step back from tools and features, and look at how we approach problems in the first place.

Starting with a simple example — the humble date field — we’ll explore how small design decisions can reveal bigger issues in how we think about users, requirements, and outcomes. From there, we’ll connect those patterns to current AI implementations.

We’ll cover:

• why “adding AI” often amplifies existing design flaws instead of solving them
• how unclear requirements lead to AI experiences that frustrate users
• practical ways to reframe problems before introducing AI capabilities
• what “good AI” actually looks like from a user and product perspective

The goal is not to discourage the use of AI, but to use it more intentionally. By focusing on the problem first, we can build solutions that genuinely help users, not just features that look impressive in demos.

This session is for anyone designing, building, or evaluating AI-enabled solutions and who wants to make sure they are solving the right problem before reaching for the technology.