There’s a predictable trap in human psychology known as the Dunning-Kruger effect: a cognitive bias where a little bit of knowledge can create a mountain of overconfidence. Innovators have a name for that first perilous peak of unearned certainty: “Mount Stupid.” The view from up there is breathtaking. For a moment, you see a limitless horizon, and right now, that horizon is filled with AI. I was there myself. I stood on that summit, fell into the subsequent “Valley of Despair” when I confronted the technology’s real-world limitations, and have been working my way up the “Slope of Enlightenment” ever since. Nearly all of Preppr.ai is built to overcome, to the degree possible, the shortcomings of AI (specifically large language models) that I learned of in the Valley of Despair.

We’ve climbed this mountain before. The parallels between AI's arrival and the internet's debut are undeniable. Recently, leaders in our field have toed into AI and correctly remember the initial skepticism and the eventual, total integration of the web. But that story is often told as a simple victory march. It wasn’t. It conveniently skips the dark sequel that began as soon as the victory was declared. To understand the stakes of AI, we must first trace the compounding crises that the post-2010 internet unleashed upon us.

How We Got Here: The Internet's Compounding Crises

By 2010, the internet's "Web 2.0" dream was fully realized. This peak optimism, however, was the starting point for a series of cascading, negative evolutions.

Phase 1: The Personal Crisis of the "Always-On" Culture

The first bill to come due was personal. The very tools designed for connection began to isolate us, while the business model of monetizing attention fostered a culture of permanent distraction.

• Algorithmic Anxiety: Social media feeds, engineered for maximum engagement, quickly became engines of social comparison, anxiety, and depression, fueling a documented mental health crisis.

• The Burnout Machine: The smartphone erased the boundary between work and life. The expectation of being constantly available led to professional burnout on an unprecedented scale, turning the promise of flexible work into the reality of inescapable labor.

Phase 2: The Societal Crisis of Weaponized Information

The platforms damaging our personal well-being soon became the battleground for our collective reality.

• From Misinformation to Disinformation: What started as organic "fake news" was systematically weaponized by state and non-state actors to disrupt elections, sow social division, and erode trust in bedrock institutions.

• The Collapse of Shared Reality: As users were algorithmically sorted into polarized echo chambers, the concept of a shared set of facts began to crumble, making good-faith debate nearly impossible.

Phase 3: The Physical Crisis of Infrastructure Attacks

With society now fully dependent on internet connectivity, the digital threat jumped from screens into the physical world.

• The Rise of Ransomware: Malicious actors began targeting the core functions of society. Hospitals were forced to turn away ambulances, 911 call centers were taken offline, and municipal governments were paralyzed.

• The Weapon Becomes the Target: The very connectivity we hailed as revolutionary became the primary vector of attack against our most critical infrastructure.

Phase 4 (Now): The AI Threat Accelerator

Today, we are entering a new phase where Artificial Intelligence is not just adding a new problem but acting as a powerful multiplier for all the previous ones. It is pouring gasoline on the fires that have been burning for the last 15 years.

AI makes misinformation cheaper and more convincing, makes cyberattacks more adaptive and potent, and can create even more addictive algorithms, deepening the mental health crisis. It takes all our existing problems and puts them into overdrive.

AI: Not Just a Better Internet, but a Different Beast

This history lesson is critical, because to compare today's AI to the early internet is to compare a fast car to a teleportation device. The order-of-magnitude change in capability brings an order-of-magnitude change in risk. The internet connected people to information; AI generates new information and can make decisions.

This creates entirely new problems:

• If you worried about algorithmic bias in social media feeds, wait until a biased AI model for damage assessment consistently overlooks low-income neighborhoods, systematically denying them resources.

• Most critically, the internet's failures are often transparent. A website is down. AI's failures can be silent and inscrutable. When a "black box" algorithm contributes to a catastrophic failure, who is responsible?

The Path Forward: Wise Adoption, Not Blind Adoption

I am not an AI pessimist. My work is dedicated to harnessing its power for public safety. But our industry's defining characteristic must be responsibility, not speed. The Silicon Valley mantra of "move fast and break things" is unacceptable when lives are on the line.

The goal isn't to be the "early adopter" but the wise adopter. The vision isn’t an AI that replaces a 911 dispatcher, but an AI co-pilot that can instantly transcribe a chaotic call and pull up building schematics while the human operator focuses on the person in crisis.

This brings us back to the Dunning-Kruger effect. Many in our field, sometimes influential leaders, are just now standing atop that first peak of confidence. This is a natural stage. I was there myself. Their enthusiasm is useful, but it must be guided past the summit of overconfidence and into the valley of reality, where real limitations become visible.

This requires a new playbook:

1. Demand Radical Transparency. We must reject "black box" solutions. If a vendor can't explain how their AI works in plain language, we can't trust it.

2. Build Ethical Guardrails First. We need clear policies on data privacy, algorithmic bias, and human oversight before these tools are ever deployed.

3. Prioritize the Human-in-the-Loop. AI should be a powerful, but fallible, advisor. The final judgment must always rest with a trained emergency manager.

The question isn't whether AI is coming. It's here. The question is whether we will repeat the past—chasing innovation from a peak of overconfidence and paying the price later—or learn from the internet's ghosts and build this new future with our eyes wide open.