Lately, I've been hearing lots of people drawing a distinction between the phrases "human in the loop" and "human in the lead,” and I’m not sure I buy it.

“Human in the loop” is an idea that's been around for a while. It's an established term that gets at the concept of meaningful human participation alongside a system like an LLM, in which the human is either in the loop or out of it. Think about an agentic coding tool that pauses to ask permission before it moves on to the next task, or a chatbot that stops to ask your opinion before it proceeds to next phase of strategic plan.

“Human in the lead,” however, is newer, and the definition is still open for interpretation. The way I think about it is that "human in the lead" describes responsibility more than participation. A human in the lead figures out what needs to be accomplished and how much autonomy to give the AI, and ultimately judges the merit of the final outcome. If you think of the classic RACI model—responsible, accountable, consulted, informed— AI may be responsible for some of the execution, but the human is always accountable for the outcome.

That's why I'm not quite convinced that “human in the loop” and “human in the lead” are actually opposing ideas. You can be both. I would also argue that the human is always in the lead, but the level of involvement changes depending on the task you’re completing. Sometimes leadership looks like being in the loop, being consulted, or maybe just being informed, but many times (always?) it means being accountable. I mean, could you imagine having an AI accomplish some incredibly useful and important task at your organization, and then, if something went wrong, saying to whoever you report to, "Sorry, I was just in the loop on that"?

In some ways, I suppose I'm splitting hairs, but the point that I want to make here is that in a world where AI can do almost any knowledge work task, you still have to make consequential decisions and stand by the results, and you need to do this from your position as the human directing the technology. You may choose to use an LLM to write for you, but you need to choose an LLM to write for you while simultaneously understanding that someone may put your writing through an AI writing detector and call you out on it. I'm not saying that's a bad thing, or that you shouldn't use AI to write, but you do need to be able to defend your ideas and explain why you made the decision that you made. The same is true for strategic plan development, debugging a codebase, or using AI to synthesize data. At the end of the day, it's your reputation on the line.

I think the real question here is not whether you are in the lead, because you are. The real question is how closely you need to lead