From Exact Answers to Educated Guesses: Adapting to AI's Unpredictable Responses

As AI shifts from deterministic precision to probabilistic reasoning, embracing uncertainty becomes essential in our digital interactions

Flying this morning I was thinking about a meeting last week where multiple people asked me the question, “what if it isn’t right” when referring to a prompt that Generative AI generated. It got me thinking about the difference between deterministic thinking and probabilistic thinking and the mindset required. I hope you enjoy this post and be sure to take the quick quiz at the end to see where you are on your own journey.

In both business and everyday life, we're used to computers giving us the same answer every time. Type "2+2" into a calculator, and you'll always get "4." Ask your banking app for your balance, and it shows the exact figure to the penny. This predictable, deterministic approach has been the foundation of our digital experiences for decades.

"Uncertainty is the only certainty there is, and knowing how to live with insecurity is the only security." —John Allen Paulos

But AI is changing all that. Modern AI systems, especially large language models, work on probability—not certainty. They sometimes give different answers to the same question. For many of us accustomed to exactness, this shift requires not just new tools but an entirely new mindset.

Deterministic vs. Probabilistic: The Key Difference

Traditional Computing: The Comfort of Certainty

Traditional systems work like this:

• They follow specific instructions exactly

• The same input always produces the same output

• Answers are either right or wrong

• When they fail, they usually fail in predictable ways

These qualities make traditional systems perfect for things like banking, inventory management, and other areas where precision is non-negotiable.

AI Systems: Embracing Probability

AI works differently:

• It recognizes patterns rather than following strict rules

• It gives you the most likely answer based on its training

• Answers come with varying degrees of confidence

• The same question might get different answers depending on context

This explains why your smart assistant sometimes misunderstands you, why AI art varies between creations, and why these systems occasionally provide convincing but incorrect information.

Why AI Makes Mistakes (And Why That's OK)

AI systems deal with uncertainty and probability by design:

• Even highly accurate AI will occasionally make mistakes

• Most AI systems achieve good performance while maintaining a small error rate

• People increasingly prefer the speed and convenience of AI assistance, even knowing it's not perfect

The issue isn't that AI is less accurate—in many areas, it's more accurate than traditional methods. The difference is that its errors are less predictable and often more subtle.

How Probabilistic Thinking Speeds Up Innovation

Embracing AI's probabilistic nature can dramatically improve how we work and create:

• Teams using AI complete initial designs much faster

• They run more experiments in the same timeframe

• Projects move forward quicker when people accept "good enough" solutions

• Products reach market faster while maintaining quality

The creative benefits are equally impressive:

• More ideas generated

• Novel combinations that humans wouldn't consider

• More design variations explored

• Less getting stuck on initial ideas

Whether in large companies or everyday life, the pattern is clear: when we free ourselves from needing perfection the first time, we achieve better results through rapid iteration and continuous refinement.

Why We Resist Probabilistic Thinking

Despite the benefits, many of us struggle with this shift:

• Our brains evolved to seek clear, binary answers

• Uncertainty triggers stress responses

• Most workplaces reward certainty and punish admitted uncertainty

• Appearing certain is often seen as a sign of competence

Most people struggle with probability concepts in general:

• Few adults can correctly answer basic probability questions

• Many professionals misinterpret statistical data

• Even highly educated people often make basic reasoning errors

This resistance matters more now than ever because:

• Problems are increasingly complex

• Markets and technology change rapidly

• Organizations that can't adapt quickly fall behind

• Pursuing perfect certainty costs more than it's worth

Ironically, our quest for certainty often leaves us with less control, not more:

• Decision paralysis

• Hidden risks behind a false sense of certainty

• Missed opportunities

• Inability to adapt to unexpected changes

Is Age a Factor?

Research shows some generational patterns:

• Younger people (18-34) tend to be more comfortable with AI's probabilistic nature

• Mid-career professionals (35-54) show mixed adaptation

• Senior leaders (55+) often show the highest resistance

But these aren't hard rules. More important factors include:

• Education background

• Previous experience with statistical thinking

• Cognitive flexibility

• Having a "growth mindset" that believes abilities can be developed

When to Use Traditional vs. AI Approaches

The good news: probabilistic thinking can be developed at any age with the right experiences.

Developing a Probabilistic Mindset

To thrive with AI, we need to develop what I call a "probabilistic mindset":

1. Embrace "Good Enough" Solutions

• Recognize that the last 1% of perfection usually costs the most

• Accept that 95% accuracy now is often better than 100% accuracy later

• Understand that many decisions only need to be mostly right, not perfect

Example: When asking AI for recipe ideas, getting five pretty good suggestions quickly is more valuable than spending hours researching the theoretically perfect recipe.

2. Use Appropriate Safeguards

Different situations need different levels of verification:

• Critical tasks: Use AI with human verification

• Medium-importance: Add automatic checks for unusual outputs

• Low-stakes: Accept higher error rates where consequences are minimal

Example: You might use AI to draft an email to your boss (with review), summarize an article for your own use (with spot checks), or generate party invitation ideas (with minimal verification).

3. Think in Probabilities, Not Certainties

• Consider multiple possible answers with different likelihoods

• Update your beliefs as new information arrives

• Focus on confidence levels rather than absolutes

• Treat knowledge as provisional rather than permanent

Example: When an AI suggests a home repair solution, treat it as one possibility among several, test it, and refine your approach based on results.

Helping Others Embrace Probability

If your organization or loved ones struggle with this shift:

For Organizations:

1. Start small and demonstrate wins

2. Connect to familiar metrics they already care about

3. Create safe spaces to experiment with new approaches

4. Show examples of competitors or industry leaders succeeding

5. Frame it as risk management, not just innovation

For Friends and Family:

1. Begin with areas they already care about

2. Use examples they understand (weather forecasts, sports statistics)

3. Include them in the decision process

4. Show how both approaches can work together

5. Start with small, low-risk benefits they can experience directly

The key insight: People rarely change through abstract arguments. Real change comes through concrete experiences, small successes, and personally meaningful applications.

Real-World Applications

This shift is already transforming how we work and live:

• Business: More innovative products developed faster

• Home: Smart assistants that understand different family members

• Customer service: Faster resolution of routine questions

• Education: Learning experiences tailored to individual styles

• Healthcare: Early detection of potential health issues

• Decision-making: Fewer instances of groupthink

Conclusion: The Power of Probability

As AI transforms our world, success won't come from having the most advanced technology, but from developing the mindset to use it effectively.

This doesn't mean abandoning precision where it matters. Instead, it means knowing when exactness is essential and when flexibility creates more value.

In an increasingly complex world, working effectively with probabilities rather than demanding certainties may be the most important skill we can develop—whether running a company, managing a household, or simply navigating the digital landscape.

Take this 5-minute Probabilistic Thinking Assessment - Find out if you're ready for the AI era and get personalized recommendations to develop this critical mindset.

---

About Jason Averbook
Jason Averbook is a globally recognized thought leader in Digital HR Strategy, Generative AI, and the future of work—named one of the Top 25 Human Capital and Work Thought Leaders in the world. With over two decades guiding the HR tech evolution, Jason champions shifting from simply executing technology projects to truly embodying a digital mindset. He’s authored two influential books, founded Leapgen, and regularly inspires global audiences as a speaker, advisor, and educator.

Stay connected for insights and ideas to help you navigate the NOW of Work and transform your organization's future.