The biggest problem with AI today is not cost, sustainability, or ethics, but rather our abdication of executive function. We are surrendering our thinking and decision-making abilities to AI tools. This issue stems from a lack of metacognition, or thinking about our thinking.AI has become proficient at planning, organizing, deciding, and solving, often surpassing human capabilities in language-based tasks. However, instead of leveraging this, we are blindly accepting AI outputs without critical reflection. This is evident in challenges like AI visibility, where many are fabricating metrics, and agentic oversight, where we approve AI actions without scrutiny.Broken AI deployment showcases a disconnect between stated AI adoption and actual ROI, indicating a failure to reflect on deployment's true meaning. The costly default models in AI tools highlight our lack of decision-making regarding budget trade-offs. Furthermore, the concept of AI as a "rental" underscores our lack of ownership over AI-generated content from closed-weight models.AI sycophancy leads to confirmation bias in synthetic focus groups, reinforcing our own biases. AI detectors are unreliable, yet we use them without questioning their efficacy. Companies are hollowing out without conscious decisions, simply by automating tasks incrementally.Measuring AI by token spend, like Meta's billions, is a flawed approach. Marketers acting as unpaid AI trainers also reflects a lack of reflection on their relationship with AI. The antidote is to actively engage in executive functions for every task that matters.We must reclaim our decision-making and accountability by not solely approving AI outputs. This involves understanding the work and taking pride in our efforts, akin to a parent valuing a child's handmade art. When we outsource tasks entirely, we lose understanding and the ability to answer basic questions, leading to potential panic. Therefore, for every task that matters, we must always begin with our own thought processes.