The Denial Mode

The second failure is the opposite: denial. You look at what AI can do and you decide it doesn't apply to you. Your work is too nuanced. Too context-dependent. Too much about human judgment. Sure, AI can write basic copy, but not the strategic work you do. It can generate code, but not the architecture decisions you make. It can analyze data, but not the insights that require real expertise.

Denial tends to lead to underestimation. It says: AI won't affect my work meaningfully. The specificity and sophistication of what I do is beyond machine capability. This mode feels safer. It feels like confidence in your expertise. But it's often a defensive story we tell ourselves because the alternative is uncomfortable.

The problem with denial is not that it's always wrong. Some work is genuinely hard for AI to do well. But denial usually means you're conflating "hard to automate" with "important" or "valuable." You might be right that AI can't replace the judgment you bring to a decision. But that doesn't mean AI won't change what you're hired to do. It might mean you stop doing the research and analysis and just do the judgment. And if judgment is what you're actually good at, great. But if you've been hiding behind the complexity of the research, you're about to find out.