Workshop: How AI Creates More Work

I would like to discuss an AI phenomenon in the world of work that has been on my mind recently, as it suggests that AI is taking us into uncharted territory in terms of how we organise our collaboration and our understanding of work, and that this presents both challenges and opportunities which we should discuss.
There is currently a lot of talk about „Workslop“ in the world of work. It is a phenomenon that arises when working with AI, and one you may well have come across yourself. The term originates from a study by BetterUp Labs and the Stanford Social Media Lab (2025)¹ and refers to AI-generated work products such as reports, presentations or summaries that look professional but lack sufficient substance to really advance a task – for example, a long report with no substance or a summary that misses the point, meaning colleagues have to rework or rewrite it. Loosely translated as „work rubbish“, which ‘clogs up’ the workflow.

It’s happened to me too: I once sent the team a 40-page analysis I’d produced using AI in preparation for a meeting. It was so lengthy that it completely overwhelmed us as a group during the meeting. That wouldn’t have happened to me in the past, because it used to take me days to write 40 pages.

Workshop – When AI leads to jobs being shifted to others

Staff members in the organisations I work with report that the applications, reports and analyses they receive are getting longer and longer, or that meeting summaries and strategy papers are circulating internally that sound so generic and vague that their purpose and message remain unclear even after reading them several times.
During a SOCIUS workshop on how to shape the AI transformation within organisations, one participant reported that, in his sector, the volume of tenders and quotations generated by AI has increased so significantly, whilst at the same time their content and intent have become so vague, that communication with project partners is reaching its limits.
In this way, the unregulated use of AI at work becomes a problem for organisations. Those who use AI to produce results but do not check them thoroughly may save time, but others are then left with extra mental work: they have to go back and check, ask for clarification or redo the work. The workload is simply shifted elsewhere.
A study by BetterUp Labs on Workslop (in collaboration with the Stanford Social Media Lab, 2025) sought to quantify the consequences. To give an idea of the scale: 40 % of those surveyed had received Workslop in the last month. Each instance took nearly two hours to correct or rewrite. It also became clear that those who send ‘Workslop’ are perceived as less creative, capable and trustworthy, and 32 % are less keen to work with that person afterwards. ‘Workslop’ therefore reduces effectiveness and trust within the team.

What teams and management can do about it

Some things can be agreed upon quickly. Teams can agree to briefly indicate in texts whether and how AI was involved, for example, „Draft created using AI, content checked“. This helps recipients gauge how much checking work lies ahead of them. It’s just as helpful to be able to reply: „That’s too long for me. What do you need from me?“, without it coming across as an affront. Above all, it’s important to talk about this and to understand each other’s intentions and how they affect others. Such agreements should be developed and can be helpful. But what particularly concerns me is that ‘workslop’ is often just a symptom of deeper-rooted issues.

What Workslop reveals about our work, and what we can learn from it

In an interview with Anne Will (12 June 2026) on the subject of AI, Aljoscha Burchardt (Deputy Site Director of the German Research Centre for Artificial Intelligence) put forward a theory which I find very constructive for team and organisational development: AI unleashes its disruptive power particularly in areas where problems already existed: systems and structures that did not function well in the past, processes that were inefficient, rules that were flouted, and a culture and values that were already at odds with one another. He said that AI radically highlights our weaknesses and gives us the chance to develop in these areas.

In this case, AI shows us where work was already poorly organised to begin with. Ethan Mollick, a professor at the Wharton School and an influential voice on the subject of AI, makes an interesting point about Workslop (interview, PwC Germany, Executive Insights, 9 March 2026): Workslop is not an AI problem, but a leadership and cultural problem. According to Mollick, the content of Workslop is often good. It is clearly worded and logically coherent. What it lacks is something else: it fails to convey a clear idea of what purpose it is intended to serve:

• After reading it, you’re not sure what’s expected of you. Are you supposed to make a decision, add to it, agree with it, or simply take note of it?
• They have a clear structure, but no discernible stance.
• The wording is so general that the context remains unclear.
• The text comes across as too polished; there is no sign of uncertainty or points of contention.

My thoughts on this: if the person giving the prompts does not clarify these points for themselves, this will not be reflected in the outcome. And if the process of developing the work is then reduced to a minimum, the scope for self-clarification also diminishes.

How is work measured and rewarded?

Mollick essentially says that such reports existed in the past too; they were simply more laborious to produce. However, as leadership has long been measured by how much work is done (hours, team size, output), there is also an incentive for staff to deliver a lot. There is a very human motive behind this: we produce output in order to emphasise our own value in line with the organisation’s internal logic. Anyone who now uses AI to ramp up efficiency is thereby exacerbating a cultural problem. 

The more fundamental question that every organisation can answer for itself is how it recognises good work. If it is not the output but rather the impact (outcome) that the work generates, it is important to define this more precisely in future – both within the team and on an individual level. This would present a great opportunity.

It might also be worthwhile to reflect on our own culture of responsibility. As we deploy AI, we need to shape and develop this culture more consciously. For one of the main challenges facing organisations – as well as society and education – is: How can people remain accountable, despite the acceleration and automation of processes through agentisation, despite the proliferation of data volumes, and despite the rise in the number of individual decisions we have to make in everyday life but which we can no longer fully understand? 

The sociologist Hartmut Rosa has described this challenge in his new book *Situation and Constellation* (Suhrkamp, 2026) as follows: as individuals, we are experiencing situations ever less frequently in which our creativity, intuition and judgement are called upon and shaped; instead, we increasingly find ourselves in rigid, rule-bound constellations that are abstract, cannot be physically experienced, and are algorithmically predetermined – just as AI-driven work and its outcomes sometimes feel. At the same time, the following applies: If AI makes it so effortless for individuals to act and produce something – be it a document, a piece of software, a process, a film or a statement – we must expand our sense of responsibility even further – taking even greater account of the boundaries and needs of others, and the time and resources of those affected by our work.

Understanding and experimenting

Workslop is just one of many phenomena arising from the use of AI in organisations, and it hints at just how much the way we work could change. Mollick argues that we need to develop the ability to experiment with new forms of work organisation. This requires courage and a spirit of discovery, because we are venturing into uncharted territory.

Let’s give it a go together

If you’d like to explore how to use Workslop as a team, we offer an AI lab for this purpose, which takes place on 10 December, from 4 pm to 8 pm online.

If you want to shape your AI transformation more consciously, whilst also breaking free from old patterns and further developing your structures: we offer government-funded INQA consultancy services for this purpose.