Compression overload is the hidden risk nobody is talking about enough in public sector AI adoption. In this episode, I explore what happens when AI compresses work so fast that human oversight can’t keep up — and what organisations need to do about it.
What Is Compression Overload?
Everyone talks about AI giving us time back. But what if it doesn’t? What if AI compresses work so dramatically that every hour saved is immediately filled with more outputs, more decisions and more actions than any person or organisation can safely absorb?
That’s compression overload. It happens when AI accelerates the movement from information to analysis, from analysis to decision, and from decision to execution — while human capacity stays largely the same. AI expands output. Human attention remains finite. That mismatch is the problem.
The Three Levels Where This Breaks Down
I look at this across three levels in the episode:
- Individual level: One person receives more analysis, options and completed work than they can meaningfully inspect — that’s cognitive overload.
- Management level: The manager is still formally approving work, but the volume and velocity make proper reflection increasingly difficult. The human is technically in the loop, but the loop is moving too fast.
- Organisational level: One person using AI can initiate in hours what might previously have required a team and several weeks. But the organisation around them may not move at the same speed.
What the Research Is Telling Us
A 2026 position paper titled AI Agents Push Humans Out of the Loop by researchers Margaret Mitchell, Avijit Ghosh and Samir Passi raises a serious governance question: when a human is expected to supervise an AI system, how much output can they actually review before oversight becomes a ritual rather than a real safeguard? They flag the danger of approval fatigue — where people keep clicking approve, not because they’ve checked the work, but because too much is arriving too quickly.
There’s also peer-reviewed research from Elena Buçinca, Maja Malaya and Krzysztof Gajos on cognitive forcing functions. Their study found that making people pause, think, or form an initial judgement before seeing an AI recommendation can reduce overreliance on the system. People don’t like that friction — they prefer seamless. But that discomfort may be exactly the point.
How to Spot Compression Overload in Your Organisation
- AI-assisted work is increasing in volume while time for review is shrinking
- Approval rates stay high, but questions, challenges and requests for evidence disappear
- Downstream teams spend more time correcting work or managing unexpected consequences
- A small group of people is expected to monitor an expanding number of AI tools and agent actions
What Leaders Should Do
The answer isn’t to slow down every use of AI. It’s to build decompression points into higher-consequence workflows — deliberate moments of friction that give people the space to function as real safeguards, not just rubber stamps.
If you want to build your AI literacy and skills, I can help
📚 Grab AI Playbook on Amazon, the perfect podcast companion.
📰 Subscribe to Human in the Loop, my LinkedIn Newsletter
🎓 Take the CPD Accredited AI Playbook Diploma
📞 Book a call at digitaltraining.ie
Sources
- AI Agents Push Humans Out of the Loop — Margaret Mitchell, Avijit Ghosh and Samir Passi, arXiv
- To Trust or to Think: Cognitive Forcing Functions Can Reduce Overreliance on AI in AI-Assisted Decision-Making — Elena Buçinca, Maja Malaya and Krzysztof Z. Gajos, ACM
- Cognitive Forcing Functions Can Reduce Overreliance on AI in AI-Assisted Decision-Making — arXiv PDF
- Article 14: Human Oversight — EU Artificial Intelligence Act
- Article 14: Human Oversight — Regulation (EU) 2024/1689
- Artificial Intelligence Risk Management Framework: Generative AI Profile — NIST
- NIST AI Risk Management Framework Playbook
Key Statistics
– An AI system can analyse dozens of documents in minutes, identify patterns, recommend courses of action and draft implementation communications.
– One person using AI can initiate in hours what might previously have required a team and several weeks.
– The position paper AI Agents Push Humans Out of the Loop was published in August 2026.
About the Podcast Host
AI and Communications Specialist | Author | Podcaster | International Speaker
Joanne Sweeney is a recognised leader in digital marketing and digital transformation, specialising in the practical application of AI to enhance communication strategies. The founder of Digital Training Institute, Joanne, empowers leaders and communications professionals to leverage AI responsibly and effectively, ensuring transparency, trust, and tangible results.
She completed the Harvard Kennedy School Executive Programme in AI and Public Policy this summer before launching her fourth book AI Playbook, in September. Joanne has also published two editions of Public Sector Marketing Pro, which were preceded by her first title, Social Media Under Investigation Law Enforcement and the Social Web.
Her books have become go-to resources for governments, public sector and policing agencies, as well as marketing and communications professionals, all navigating the challenges of the changing Internet.
Joanne is the host of the AI SIX Podcast, a daily six-minute podcast that provides actionable insights, expert interviews, and practical advice to help professionals navigate the new world of work.
