Inspiration

We were inspired by the dangerous digital echo chambers created by engagement-driven algorithms and AI assistants that feed users single-sided information. Rather than letting researchers unconsciously fall into confirmation bias, we wanted to build a tool that does the exact opposite of big tech: actively challenging your assumptions before you make a final decision.

What it does

Dynamic Bias-Breaker is an AI-powered research middleware and interactive web application that monitors your semantic clustering in real time. If it detects that your search queries or arguments are leaning entirely to one side of a debate, its "Devil's Advocate" engine automatically triggers to display critical blind spots, counter-evidence, and balanced alternative perspectives.

How we built it

We built a responsive, dark-mode single-page web application featuring a glassmorphism UI using HTML5, modern CSS, and vanilla JavaScript. We engineered an interactive research sandbox with multi-domain preset tracks, a real-time semantic scoring telemetry gauge, and a dynamic state machine that synthesizes and displays structured counter-arguments on demand.

Challenges we ran into

Our biggest challenge was solving a unique UX contradiction: how to design an intervention engine that interrupts biased workflows powerfully enough to matter, without causing user defensiveness or adding sluggish latency. We also had to craft reliable heuristic parsing capable of instantly evaluating diverse subjects like tech architecture, economic policy, and AI automation.

Accomplishments that we're proud of

We are proud of successfully translating a complex, abstract psychological concept into a sleek, fully working interactive prototype under tight hackathon constraints. Creating a smooth "Cognitive Interrupter" card that dynamically flips perspectives without requiring heavy external API calls was a massive technical win.

What we learned

We learned the core mechanics of cognitive bias and semantic clustering in digital research. More importantly, we gained deep insights into proactive UI design—specifically, how crucial it is to inject balancing telemetry at the point of decision rather than after a stance has already solidified.

What's next for Dynamic Bias-Breaker

We plan to scale this prototype into a fully functional browser extension and API middleware layer. Future roadmap items include integrating local, privacy-focused open-source LLMs (like Llama 3) to analyze live web browsing data in real-time and introducing deep analytical dashboards to track a user's information diet health over months.