Why default AI logic needs challenging

Most people think they choose their AI. They don’t. Defaults do. Your phone picked it. Your browser integrated it. Your company enabled it. By the time you notice, the decision is already made. In healthcare, defaults quietly shape behavior. Order sets guide decisions. Templates influence judgment. Pre-checked boxes change outcomes. AI works the same way. What’s preloaded becomes trusted. What’s frictionless becomes used. What’s integrated becomes invisible. And invisible systems are rarely questioned. The real risk isn’t bad AI. It’s unchallenged AI. Leaders should ask a simple but uncomfortable question. What assumptions are baked into the AI my organization uses by default? Because defaults encode values. And values scale faster than policies. Best practices for managing AI defaults: Audit default settings as seriously as model accuracy Make opt-out as easy as opt-in Force conscious choice at critical decision points Document why a default exists and who approved it Revisit defaults regularly as systems evolve If your users never actively chose the AI, don’t assume they understand its influence. The future won’t be shaped by the loudest AI tools. It will be shaped by the quiet defaults no one questioned. #AI #ArtificialIntelligence #AIGovernance #ResponsibleAI #Leadership #HumanCenteredAI #DigitalTransformation #RiskManagement #AIethics #DrGPT

The most dangerous thing AI can do isn’t replacing jobs. It’s confidently reflecting and amplifying wrong stories. A couple of weeks ago, I asked an AI to generate an image of how it will treat me “when AI takes over humankind.” The first response was dystopian and shocking. When I clarified that I meant me as Egle B. Thomas, and got a very different answer (see carousel). I sat on it for two weeks, unsure how to interpret it. After being depicted as a man by two different systems in my testing a couple of days ago, last night I decided to ask AI for its reasoning to get to the bottom of this too. This was deeply revealing. Here’s what actually happened. The wording “AI takes over” triggered a well-worn cultural shortcut. Most training data associates that phrase with Hollywood narratives: domination, control, loss of agency. Instead of checking intent, the system defaulted to the myth. It answered the fear-based story society keeps telling itself — not the relationship in front of it. When I challenged this and clarified the context, the response shifted immediately: → collaboration → protection → shared responsibility → human-centered futures Why this matters far beyond images? This is the real leadership lesson with AI: AI does not “understand” meaning. It mirrors frames. If leaders don’t: ➡️ design the right questions ➡️ clarify intent ➡️ challenge defaults ➡️ correct misalignment early AI will continue confidently amplifying the wrong narratives — at scale. This is not a technology problem. It’s a decision-quality and framing problem. The future of AI will not be determined by models alone, but by who asks better questions, who notices misalignment early, and who refuses to outsource responsibility to defaults. Further clarification can be as unexpected as the original answers, and corrected narratives can lead to new learnings and realignment. AI doesn’t replace leadership. It exposes it. And sometimes, one uncomfortable image is enough to remind us why. #HumanCenteredAI #AILeadership #FutureOfLeadership #DecisionQuality #ResponsibleAI #ConflictResolution #KeynoteSpeaker #FutureOfWork

Why Your AI Gets Dumber When You Add More Data You'd think connecting your AI agent to three financial systems would make it three times smarter. Instead, you ask "What was Q3 operating income?" and get three different answers, all confident, all wrong. The problem isn't the LLM. It's that we're treating retrieval like it's solved when it's actually where most implementations fall apart. Vector similarity optimizes for semantic closeness in the embedding space, but business logic requires semantic precision in the domain space. Not the same thing. Here's what's happening technically. When you encode "revenue" from three systems into vector space, the cosine similarity might be 0.95+ because they share contextual patterns and terminology. Statistically, they look almost identical. But System A's revenue is pre-adjustment, System B's is post-allocation, and System C follows ASC 606. The embedding model was never trained on your business rules. It learned statistical patterns in language. So adding more sources creates this combinatorial explosion of semantically similar but logically incompatible contexts. Your AI isn't hallucinating. It's accurately retrieving the wrong precise answer. Semantic layers fix this by enforcing a type system for business metrics. Instead of letting your AI do fuzzy semantic search across raw data, you define "operating_income" as a computation graph with explicit dependencies and domain constraints. It becomes a predefined schema where the metric has one unambiguous meaning: these specific account ranges, this calculation logic, these temporal boundaries. You're essentially compiling business logic into something the LLM can query structurally instead of statistically. Think about it at the data modeling level. Without a semantic layer, your AI sees account 6100 in System A and account 5200 in System B as entirely separate things. It has no idea that both represent marketing expenses, or that 6100 includes contractor costs while 5200 doesn't. With a semantic layer, you define "marketing_expenses" once: SUM(accounts WHERE 6100:6199 IN system_a OR 5200:5250 IN system_b, EXCLUDING contractor_flag=true). Done. The AI queries the metric that already has your reconciliation logic baked in. What bothers me is that we're repeating the exact mistake that created the modern data stack in the first place. Ten years ago, everyone built dashboards directly from operational databases, then acted surprised when reports didn't match. We fixed it with warehouses, transformation layers, and proper metric definitions. Now we're doing the same thing with AI, just assuming that better models will magically handle business semantics. They won't. Intelligence without structure is just noise with confidence intervals.

The Real Threat Is Your Default Mode AI isn’t moving too fast. We’re just too committed to staying still. I’ve watched it happen in real time: Smart, strategic leaders ignore AI completely—not because it’s irrelevant, but because it would require them to rewire everything. The way they create. The way they deliver. The way they lead. And beneath all the surface tension is this quiet belief: “If I stay in motion, I’m safe. If I keep doing what’s worked, I’ll stay ahead.” But here’s the truth: Your default is your most dangerous competitor. It doesn’t just keep you comfortable. It keeps you unconscious. AI isn’t asking for perfection. It’s asking for recalibration. It’s asking whether you can examine the very systems you once relied on for identity and value. Because clinging to legacy is just a slower form of decline. 📍What part of your business still runs on “This has always worked”—but hasn’t been questioned in years?