Shaping AI Behavior Workshop: The Values You Ship

See how product decisions become AI behavior. A two-hour, hands-on workshop for people building with AI. AI features force us to choose what counts as "helpful," "safe," or "fair”, especially when those priorities conflict. These choices aren't just theoretical; they are embedded in model selection, defaults, prompts, refusals, and product policies. Through case studies, product testing, prompt design, and red teaming, you'll examine these trade-offs and learn how to translate human values into concrete system rules. AI's Actions are Product Decisions  Terms like "helpful" and "safe" sound straightforward until they have to govern a specific interaction: These questions require judgments about the boundaries of your product. Throughout the workshop, we’ll put these questions to the test, comparing what companies say they value against how their AI actually responds. Workshop Activities  Each phase of the workshop gives you a new lens for examining and shaping AI: What You’ll Learn Who Should Attend  This workshop is for the people shaping AI products: product managers, engineers, designers, researchers, content strategists, and students entering the field. It is especially relevant to teams shipping (or considering) AI features where refusals, personalization, disclosures, or edge-cases are a core part of the product design. No ML engineering or philosophy background required. The exercises rely on prompting and discussion, not code. Basic familiarity with a tool like ChatGPT or Claude is all you need. Why an In-Person Workshop? You can read about ethical frameworks anywhere. This workshop is about applying them. You'll compare interpretations, defend design decisions, and test one another's bots in real-time. Differences in human judgment become part of the exercise, and the Red Team vs. Blue Team challenge gives you a live human opponent actively hunting for weaknesses in your specs. Logistics & Facilitators FAQ I already work on these questions. What will this add for me? A structured process for connecting high-level values to specified behaviors and test cases, plus the rare opportunity to compare your decisions with other practitioners and test them in real-time. Is this just a lecture about why AI is bad? No. We don't prescribe one "correct" set of values. This is a practical look at how product decisions are made, implemented, and tested. Will there be a recording? No. Because this is heavily interactive and discussion-based, the experience doesn't translate to video. Can my team attend together? Yes, and group rates are available. It’s highly beneficial for colleagues to work through the same cases and align on how they would handle them in their own products.