Using AI as a Sparring Partner Instead of Ghostwriter

Integrating generative AI into executive communication creates a paradox: the tools that speed up work also threaten to erode the critical thinking and empathy essential for leadership. While AI is effective at organizing data and structuring ideas, its tendency to produce average output leads to a generic corporate voice that masks a lack of original judgment. Leaders who use AI as a substitute for the writing process risk empathy atrophy and a widening comprehension gap, leaving them unable to defend or deeply understand the ideas they present. The competitive advantage in this era belongs to those who use AI as a sparring partner to sharpen their own thinking, rather than as a ghostwriter to automate their presence.

The Hidden Cost of Polished Communication

The primary danger of AI in communication is the illusion of competence. Because LLMs are designed to be goal-seeking and pleasing, they often provide what the user wants to hear rather than what is objectively true or insightful. When leaders outsource the drafting process, they bypass the operationalized critical thinking that happens during the act of writing.

> The big risk with AI from my perspective is the outsourcing of empathy. AI can be a wonderful tool for helping you enable empathy, you can think about and use AI to help you hone your message based on characteristics you provide but at the same time it can cause empathy atrophy if all we do is outsource the most challenging communication we have then we are in trouble.
>
> -- Matt Abrahams

Over time, this creates a feedback loop of anxiety. When a leader is put on the spot to defend a polished document they did not truly synthesize, they lack the depth of knowledge to respond effectively. The system rewards the speed of the output but penalizes the long-term credibility of the leader.

The Authenticity Crisis and the Averaging Machine

AI models are averaging machines. They thrive on consensus and commonality, which is the opposite of distinct leadership. When leaders rely on these tools without rigorous human intervention, they risk sounding like corporate bots.

> My colleague Scott Bernato loves to say that LLMs are really just an averaging machine so if you want to be average, you should use them.
>
> -- Allison Beard

To combat this, leaders must treat AI as a sparring partner. Wade Foster of Zapier notes that he intentionally updates his system prompts to demand critique and debate rather than passive agreement. By forcing the AI to challenge assumptions, the leader uses the tool to build a more robust argument, rather than simply smoothing over the raw edges that make a human voice authentic.

Architecting for Human Connection

Systems thinking requires us to look at how the environment dictates the quality of communication. If a leader suspects their team is relying on AI to masquerade as them in Slack or during interviews, the solution is not better detection software; it is changing the environment.

This means re-introducing blue book scenarios: spontaneous, in-person discussions where the ability to think on one's feet cannot be faked. As AI makes polished, low-effort communication common, the ability to demonstrate genuine, messy, human judgment becomes a high-value differentiator. The system responds to your inputs; if you input low-effort prompts, you receive low-effort, homogenized results that erode the team ability to think for themselves.

Key Action Items

• Audit your Empathy Threshold (Immediate): Identify high-emotion communications, such as negative performance reviews or conflict resolution, and commit to drafting these entirely without AI assistance to prevent empathy atrophy.
• Implement Sparring Partner Prompts (Over the next quarter): Update your AI system instructions to explicitly request critique, counter-arguments, and skepticism rather than just drafting assistance.
• Establish Transparency Etiquette (Immediate): Adopt a policy of declaring the level of AI involvement in your communications, such as AI-assisted research versus AI-drafted text. This builds trust and helps the reader calibrate their expectations of your personal investment.
• Connect AI to Internal Data (12-18 months): Stop using generic LLMs for organizational strategy. Hook your agents into your specific company data, such as Slack, project management tools, or internal codebases, to move from general coaching to context-aware insights.
• Practice Spontaneous Q&A (Ongoing): Use AI to role-play as a skeptic or a subject matter expert to rehearse for high-stakes meetings, but focus on the reps of answering, not memorizing the AI output.
• Prioritize Human-in-the-Loop Metrics (12-18 months): For automated communication workflows, such as customer support bots, establish statistical rigor and internal evaluations to ensure the AI is not drifting into incorrect or vague responses.