AI fatigue

As if out of nowhere, artificial intelligence is everywhere and found in everything. It is in all our electronics, from the latest laptop and software offerings to cars, washing machines and smart robotic vacuum cleaners. And it is in all of our institutions, from education to all levels of government, as federal, state and local officials wrestle with developing guiderails and guardrails for its use. Understandably, people are overwhelmed by AI’s possibilities and risks. At a recent local government CIO event, participants began expressing this sentiment through the term “AI fatigue.”

How can this be? After all, generative AI is less than three years old. So, what is AI fatigue? It refers to the exhaustion or the sense of being overwhelmed by the rapid proliferation and constant engagement with artificial intelligence technologies, including the pressures to do more and budget more. There is often a gap between the ambitious promises of AI capabilities and the actual results delivered in government applications. This mismatch can lead to skepticism and fatigue among government officials and employees.

What’s more, there is often tension between the push to adopt cutting-edge AI technologies and the need to address more immediate, practical challenges in government IT operations. Scaling AI solutions within government organizations also face significant hurdles, which include high costs, risks associated with implementing a new technology, difficulties in data preparation and management, challenges in redesigning processes and roles to incorporate AI effectively, rapid evolvement of newer AI applications, and staff training on AI usage and experimentation.

These complexities all contribute to AI fatigue, a phenomenon that is manifesting itself in several ways. Here are a few examples:

Cognitive Overload. Individuals may experience mental exhaustion from the continuous interaction with AI tools, whether through chatbots, recommendation systems or virtual assistants. The frequent need to adapt to new AI-driven technologies can lead to cognitive strain.

Technological Overload. Organizations or individuals may feel overwhelmed by the sheer volume of AI tools, platforms and updates, making it challenging to keep up with the pace of change. This can lead to decision paralysis or resistance to adopting new technologies.

Emotional or Ethical Concerns. Repeated exposure to AI-driven systems, particularly those that make decisions or predictions about human behavior, can lead to discomfort, ethical concerns or distrust. This can cause individuals to become weary or skeptical of AI technologies.

Burnout in Development and Management. Professionals working in AI development or those responsible for managing AI systems may experience burnout due to the high demands and fast-paced nature of the field. The pressure to continuously innovate and the complexity of AI projects can contribute to fatigue.

Fear of job displacement. Many government workers are concerned that AI could make their jobs obsolete or replace certain job functions.

Increased workload expectations. Employees worry that AI implementation will lead to increased productivity expectations and workloads without additional resources.

Rapid pace of change. The fast evolution of AI technology is causing anxiety.

Information overload. The sheer volume of information about AI capabilities and applications can overwhelm government decision-makers trying to separate hype from genuine opportunities.

Resource constraints. Many government agencies are not getting additional resources to properly evaluate and implement AI, straining already limited resources

AI fatigue raises two essential questions: why is this happening now, and what can be done to alleviate its symptoms?

Why Now? Let’s Start with the First Question. Why Now?

Considering generative AI only entered the market in late 2022, what makes this emerging technology more worrisome than cloud computing, broadband proliferation and eGovernment/digital government services? The term fatigue never appeared when it came to other emerging technologies.

The answers appear to reside with the speed in which AI and generative AI took hold for both personal and professional use. It only took five days for ChatGPT to reach more than 1 million users, compared to the five years it took for Netflix to reach 1 million subscribers.

Another underlying reason is that AI is a technology that emulates human productivity and creativity. AI can supplement almost any repetitive business process with near-unimaginable speed.  Some have commented it takes far longer to type in a prompt to a generative AI client than it takes to receive a response—often in fractions of a second. AI generates information so quickly that it can be unnerving.

IT managers also worry that any AI-generated output could be flawed or dead wrong and ultimately could cause unintended harm to citizens. Here, data quality and integrity are essential, and IT managers rightfully worry that their data is clean and unbiased.

Finally, there are concerns about AI being out of IT management's control and questioning the integrity of the underlying algorithms.

In summary, AI is unlike any other emerging technology in history, and it continues to grow exponentially, providing IT leaders with less time to think and plan. As quickly as this new technology takes hold, IT managers must deal with its many uncertainties.

The Remedy

IT managers and users need to recognize that everyone is feeling the same, everyone has the same concerns and fears. But there is safety in numbers. AI-induced stress can be alleviated by human social interaction and seeking out user groups and forums for moral support, leadership, experimentation and sharing of best practices.  

One need not look far to find great support organizations, such as my own, the Public Technology Institute, which has featured numerous AI sessions and writings and will soon publish the second annual City & County AI Survey Report. We also recently published the book Artificial Intelligence—A Primer for State and Local Governments.  

Then there is the city of San Jose’s Government AI Coalition, which offers the largest network of AI government users in the U.S., as well as resources and policy guidelines. Of course, many other organizations at the state and national levels provide AI information, research and networking opportunities that can be helpful. In addition, many community colleges and universities have formed AI centers of excellence that always want to collaborate with state and local governments. Each government unit should consider creating an “AI community of interest” to share the latest AI-related experiences. The point is that with technology such as artificial intelligence, IT leaders face new challenges every day, making it imperative to reach out and join the many excellent peer networks for practical advice and even a bit of handholding.

To address these concerns, government agencies must focus on gradual AI implementation, provide comprehensive training, involve employees in the adoption process, and emphasize how AI can augment rather than replace human workers. Clear communication about AI's role and ethical guidelines is crucial to alleviate employee anxiety.

One must remember that while artificial intelligence has been around for more than 50 years, its offspring, generative AI, is still in its infancy. At the age of 3, it has already learned to speak and converse with humans. Unlike human infants, AI never sleeps or naps, increasing stress and fatigue. As generative AI grows, IT leaders must recognize AI fatigue and its symptoms and understand that human networks will provide the best relief.

Dr. Alan R. Shark is the executive director of the Public Technology Institute (PTI) and associate professor for the Schar School of Policy and Government, George Mason University, where he is also an affiliate faculty member at the Center for Advancing Human-Machine Partnership (CAHMP). Shark is a National Academy of Public Administration fellow and co-chair of the Standing Panel on Technology Leadership. Shark also hosts the bi-monthly podcast Sharkbytes.net. Dr. Shark acknowledges collaboration with generative AI in developing certain materials.