AI can reduce the time needed to draft, summarise and organise work. But for some employees, especially people drawn to novelty or managing ADHD, the same tools can create a new form of distraction: an apparently useful task with no natural stopping point.
Peter Shankman frames this risk directly in his essay, “AI Will Happily Eat Your Whole Afternoon.” His central warning is not that AI has no workplace value. It is that a system designed to offer another prompt, another option or another refinement can become an “infinite optimisation machine” for people who are already susceptible to chasing the next interesting possibility.
Peter Shankman on ADHD, AI and the productivity trap
The productivity trap begins when a task that should be finite becomes open-ended. A team member may start by asking AI to improve an email, then request a different tone, more alternatives, a shorter version, a longer version, a campaign outline and a set of follow-up messages. Each individual prompt can feel productive. Taken together, they can consume the time intended for judgement, decisions and delivery.
For people with ADHD or a strong novelty-seeking tendency, this pattern may be particularly compelling. The issue is not a lack of capability or effort. Rather, the abundance of instant options can make it harder to decide when the work is sufficiently good and when to move on.
“AI Will Happily Eat Your Whole Afternoon.”
Shankman’s observation offers a useful corrective to a common adoption question. Instead of asking only, “What can this tool do?”, leaders can also ask, “Which part of this task actually requires the tool, and what signals that the task is finished?”
Where AI can help—and where boundaries matter
AI can be useful when the desired output and the limit are clear. For example, it may help a manager turn rough notes into a first draft, generate discussion questions for a workshop or summarise a long document before human review. In these cases, the tool supports a defined next step rather than replacing responsibility for the work.
It can become less helpful when the activity has no agreed purpose, owner or endpoint. Leaders and L&D teams can reduce that risk by building practical boundaries into AI use:
- Define the task before opening the tool: draft, critique, summarise or generate options.
- Set a time limit for experimentation, particularly for open-ended creative work.
- Specify a decision rule, such as selecting one draft after two iterations.
- Keep human review for material that affects clients, employees, reputation or policy.
- Invite employees to discuss which workflows help them focus and which create friction.
These are not restrictions on curiosity. They are ways to ensure that curiosity serves the work rather than quietly displacing it. They also recognise that a single AI workflow will not suit every employee equally.
AI can assist communication, but it cannot create trust
Shankman’s perspective also matters for customer experience and leadership communication. AI can help structure a message or prepare a first draft, but it cannot take responsibility for what is said, understand the full relationship context or demonstrate genuine care on a leader’s behalf.
Trust is built through accountable human behaviour: listening, responding appropriately, being clear about expectations and following through. An AI-generated message may be polished, but polish alone does not make it personal, accurate or credible. That distinction is important for marketing teams, customer-facing leaders and managers communicating with employees during change.
A practical question for organisations is therefore not whether AI should appear in communications at all. It is whether the human sender has added the judgement, specificity and ownership that the situation requires.
A neurodiversity-aware approach to AI adoption
AI implementation is often discussed in terms of features, access and efficiency. Shankman’s argument adds another consideration: attention. HR and technology leaders can ask whether a new tool simplifies a workflow or merely adds more places to explore, edit and second-guess.
That means involving employees in the design of AI-enabled processes, allowing for different working styles and making boundaries explicit rather than assuming that everyone will self-regulate in the same way. A tool can be optional for ideation, for instance, while standards for approval, confidentiality and final accountability remain clear for everyone.
What audiences can take from a Peter Shankman session
For leadership, HR, L&D, customer experience and communications events, Shankman’s lived-experience lens can prompt a more grounded conversation about responsible AI use. Rather than treating productivity as a simple output metric, audiences can examine how attention, novelty, workflow design and authentic communication shape the day-to-day experience of work.
His perspective is particularly relevant to programmes on neurodiversity at work, AI adoption, manager capability and customer trust. The useful starting point is simple: deploy AI deliberately, set boundaries around its use and preserve the human judgement that relationships and decisions demand.
To enquire about Peter Shankman for a conference, leadership meeting or corporate event, contact Speakers Connect at info@speakersconnect.com.