Tag: Wicked problems

Can AI Help Us Choose a Better Path?

On By gyonas3 Comments

We live in a world haunted by wicked problems—problems that mutate, multiply, and resist every attempt at resolution. Climate collapse. Authoritarian drift. AI accelerating beyond oversight. Debt spirals. War without end. Each of these crises feeds on the others, creating a Gordian knot that neither our leaders nor our institutions seem able to untangle.

What can be done?

One idea, admittedly audacious, is to take a lesson from history. In 1953, President Eisenhower launched Project Solarium, a bold experiment in structured strategic debate. Three teams of experts—each representing a different philosophy toward Soviet containment—were asked to develop, argue, and defend their strategies. Eisenhower didn’t pick favorites; he wanted the sharpest disagreements and clearest thinking to shape America’s Cold War policy. The outcome helped solidify the doctrine of containment, which guided U.S. policy for decades.

Could we do something similar today—only smarter, faster, and more inclusive?

That’s the premise of a Solarium 2.0. Not a political stunt or top-down policy declaration, but a structured process to think our way through global dilemmas. Teams of diverse thinkers—equipped with AI-enhanced reasoning tools (software that helps analyze data, challenge assumptions, and model outcomes in real-time)—would be guided by an experimental new concept: a values-based AI coach. This coach, rather than dictating answers, prompts participants to consider ethical frames, recognize cognitive bias, and evaluate decisions through shared human values like justice, sustainability, and long-term resilience. Why a coach? Because even smart, ethical people face cognitive overload, fall into groupthink, or overlook long-term consequences—especially under stress. A values-based AI coach doesn’t replace human judgment; it enhances it by prompting reflection, highlighting ethical dimensions, and encouraging diverse viewpoints in real time.

It’s part war game (testing how different strategies play out under pressure), part design studio (developing novel solutions), and part constitutional convention (re-examining foundational assumptions and frameworks for action). The goal isn’t to draft legal documents but to explore better ways to think, act, and collaborate in a world of accelerating complexity.

But wait—who gets to play? Who chooses the questions? Why should anyone listen?

These are critical questions. And they lead to a bigger one: Should the new Solarium be national or global?

Solarium 2.0 is envisioned as an international initiative, drawing participants from multiple countries and cultural perspectives. Today’s decision-makers—men like Putin, Trump, Netanyahu—often play zero-sum games. But our survival may depend on win/win solutions. We may not have an Eisenhower today. But we can build the process he pioneered—updated for our era.

AI-assisted collective intelligence—humans working with AI systems to make better group decisions—isn’t science fiction. It’s a growing research field. But most systems today focus on narrow tasks or post-hoc analysis. What we need is a leap forward: real-time human-AI decision-making for wicked problems, with the AI acting not as oracle, but as orchestrator.

In a 2024 review of human-AI collaboration published in Intelligent-Based Systems, Hao Cui—a leading researcher in multi-agent systems—surveyed dozens of experiments. Most involved controlled environments and constrained tasks. But none attempted a live simulation of high-stakes, global-scale problem-solving under cognitive load. Why not? Possibly because of technical hurdles—or perhaps because we’ve lacked the will to convene the right people with the right tools.

That’s why I’m calling for a bold experiment: the Solarium 2.0 War Game.

Convene it under the auspices of a respected neutral body—such as the U.S. National Academies of Sciences, Engineering, and Medicine—and draw participants from national labs, forward-thinking companies, and leading universities. Equip them with emerging AI platforms designed not just for speed or scale, but for principled, values-aligned decision support. Use real-world scenarios, real time constraints, and real disagreement.

We’ve theorized enough. Now it’s time to find out:

Can we do better—together?

(Disclosure: This document represents a collaboration of human and artificial intelligence. The illustration was created by AI.)

Gerry in garage

Again with the hereafter

On By gyonasLeave a comment

As I mentioned in an earlier blog post, now that I’m in my 80s, I frequently find myself dwelling on the hereafter. In other words, I find myself walking into the garage in the middle of an errand, presumably fetching something for my wife, and I ask myself, “What am I here after?” I’ve even tried asking AI, since most people are outsourcing tasks to AI nowadays. Alas, AI (like me) had not listened to my wife.

This leads me to consider AI’s shortcomings (a much better choice than dwelling on my own shortcomings!) But, does AI have shortcomings? As artificial intelligence continues to accelerate in speed and scale, many assume that the future of problem-solving will be machine-dominated. But what if neither AI nor unaided human reasoning is sufficient to address the wicked problems of our age. Wicked problems have no closed end solutions. They defy simple answers and cut across traditional domains of expertise. They require not just faster computation but deeper wisdom—and new ways of thinking.

Before turning to AI for assistance, let’s start by considering what we can do to make people think better. Is there a way to boost human intelligence and cognitive abilities? If you have read my latest novel, The Dragon’s Brain, you are familiar with an invention called the Brainaid, a form of non-invasive brain stimulation and brain wave entrainment that improves focus, creativity, memory, and emotional regulation. If you haven’t read The Dragon’s Brain, what are you waiting for? Head to your favorite bookstore or website and get your copy now. (Here’s the Amazon link.) In any case, the fictional Brainaid is based on a similar brain stimulation device that I patented many years ago. While a fully functional product is yet to hit the market, I firmly believe that science can–and will–enhance human thought. AI is getting smarter, so why not enhance the human side of the equation too?

The current narrative often positions AI as a replacement for human intelligence. But the reality is more nuanced. Machines excel at speed, scale, and pattern recognition. Humans contribute meaning, ethics, empathy, and insight. Each is powerful on its own—but neither is complete.

AI alone can be dangerous. It lacks human intuition and empathy. Perhaps you’ve met a physicist, engineer, or computer scientist who also lacks human intuition and empathy. I earned my bachelor’s degree in engineering physics from Cornell University, where I was trained to think precisely, quantitatively, and rigorously. But two courses outside the engineering curriculum had an even greater long-term impact on how I think. The first was a course in the Philosophy of Science taught by Max Black, a sharp and elegant thinker who challenged us to ask not just how science works, but what we mean when we claim to know something. The second was a course in the Philosophy of Religion taught by Milton Konvitz, who opened my mind to the moral foundations of law, liberty, and human dignity—drawing from both secular and religious traditions. These classes taught me to ask hard questions, tackle wicked problems, and never separate the technically possible from the ethically responsible.

That’s why I propose a purposeful collaboration between humans and AI. There is no need to hand our cognitive duties over to machines, rather, we need to enhance our abilities and learn to use artificial intelligence to help us tackle the wicked problems we’ve been unable to solve.

Imagine a decision system composed of diverse humans enhanced for clarity, openness, and ethical discernment, using AI to optimize data processing, information retrieval, and scenario modeling, Imagine humans working in partnership with AI to develop a problem-solving approach that supports feedback, deliberation, and adaptation with accountability, transparency, and value alignment. This is not a hive mind. It is more like a meta-cortex—a layered neural system in which individual insight is preserved, but multiplied through structured collaboration and augmented intelligence.

If we succeed, we may find ourselves entering a new age: not of superintelligence, but of super-collective wisdom. We may experience:
-A future where leadership is not about who dominates, but about who understands,
-A future where advancing age, like mine—95 in a decade—is not a barrier but an asset,
-A future where the tools of science and technology serve a higher purpose: to help us decide wisely, together.

So, as I stand in the garage asking myself., “What am I here after?”  Maybe just that.

Terrorism is a wicked problem

On By gyonasLeave a comment

Seventeen years ago, I was invited to give a presentation to a group of intelligence analysts at the Joint Military Intelligence College. At that time, I was managing the Advanced Concepts Group at Sandia National Labs, and my group was focusing much of our attention on emerging threats. A current issue was what was called “the global war on terrorism.” This war began in Afghanistan in 2001 after the Al-Qaeda attack and continued for 20 years. During that period, it expanded to include Iraq in 2003 with one justification being the belief that Iraq was linked to Al-Qaeda.

The threat of terrorism was very much a major national security issue, and my presentation attempted to address these issues based on my perceived needs for intelligence analysis. I had been increasingly interested in dealing with current complex challenges, and I studied the literature of systems engineering approach to solving relevant problems. What I learned was that most of my career as an engineer and physicist had been dominated by what were called tame problems, and the national security issues of the time were best described as wicked problems that would be long in duration.

I was convinced that the current military issues were best described by a timeline beginning with a long period of increasing threat, and short period of conflict, and a much longer post conflict period of managing the threat. I thought the key to success of this challenge would be in the hands of intelligence analysts that knew how to deal with wicked not tame problems.

Tame problems had been the focus of my training and career, and are the typical challenge for analysts, engineers, and convergent thinkers. Such problems have a well-defined problem statement. For example, a tame problem is figuring out how to build a bridge. A wicked problem includes planning for the bridge, obtaining permission from the community and elected officials to build it, acquiring funding and scheduling, and working with the various individuals and agencies required to build that bridge. The bridge builders know what, where, and how to proceed with a well-defined end point of the task. They can learn from the records of other similar bridges already built and can easily try out various paper designs and choose the one most appropriate approach. They have an orderly approach to analysis, design, and implementation, but do not have the divergent thinking approach that is actually needed to complete all of the tasks involved in building the bridge.

The typical tame approach can lead to disaster if the problem is really wicked. If the problem is defined incompletely, prematurely, or influenced by desperation, ambition, fear, greed, hatred, or other emotions. Being driven to a hurried solution can lead to oversimplified solution options and an early and false belief that the problem is solved. The different perspectives, backgrounds organizations, and prejudices can lead to escalating confusion, conflicts, and paralysis.

A symptom of a wicked problem that is treated as tame is when the leader says, “Let’s get organized, put the right person in charge, get on with the solution, and get it done.”

The tame problem approach is a satisfying and coherent method of increasing knowledge.  Wicked problem solving, on the other hand, can often be characterized by frustrating alternating periods of euphoria and utter depression. So, are wicked problems just another worthless activity that is in the end a hopeless mess?

Well, maybe, but if you know the problem is truly wicked, a wicked engineering analyst can make real progress by spending a great deal of time and effort to comprehensively formulate the approach as a nonlinear spiral instead of a ladder of subsequent steps. The key is also to share the complexity with a group of creative thinkers and communicators that have a diversity of views. It’s important to share ideas frequently as the context of system issues changes and avoiding a focus on the detailed piece of the problems. Since premature belief in success will turn out to be the devil preventing group productive cooperation, the participants need to trust each other as the game changes.

I concluded that without active counterterrorism intervention, the level of terrorist violence will be low until a triggering threshold is passed, and, at some point, conflict will demand increased security emphasis.  If successful, counterterrorism actions can be taken that will lead to a cessation of combat operations. This period will be followed by a long period of stabilization and reconstruction. During the active combat period, the adversary may apply such irregular methods as assassination of leaders; hostage taking; cyber, bio chem, and infrastructure attack. The adversary may also introduce social and psychological methods such as induced chaos, exponential migration, financial attack, and race wars. The symptom of terrorist success would be a disruption of societal stability and stimulation of self-destructive behaviors.

A strength of the wicked engineering group could be the application of ubiquitous information technology, but in the hands of the terrorist, could also accelerate instabilities, so it will be necessary to take advantage of advances in complex computer modeling and simulation as well as the application of neuroscience to enhance cognition and group problem solving.

By gaining a neuro advantage over the adversary, methods of deterrence and dissuasion will become apparent. The advances in the neuro science spectrum can enhance the psychological armor, accelerate learning, cognition, and memory. Use of such methods applied against the terrorist can create confusion, fear, and loss of understanding of the rapidly changing environment. The adversary’s use of such psychological and information warfighting tools can lead to our early failure in dealing with the threat. The positive and negative implications must be understood.

At the time of my presentation, I believed that neuroscience advances in the hands of the adversary (which won’t have the same legal and moral constraints that we have) would have an important impact on the outcome. I also believed that the challenge of dealing with terrorism was open ended and there would likely never be a last move in this contest, so the happy ending to the story was not obvious.

My new novel, The Dragon’s C.L.A.W., also tackles a wicked problem. The protagonist realizes that his breakthrough invention, which has the ability to transform the world by providing clean, affordable, unlimited energy, can also be used to create a deadly weapon. I called upon my understanding of how to wrestle with wicked problems as I described how the character dealt with the conflict the dual nature of his work. Wondering how he resolves the problem? You’ll have to read the book!

Have you ever dealt with a wicked problem? How did you approach it? Did you resolve it? Comment below.