Dependence, Control, and the Coming Age of AI Reliance
Date Started: 10/19/2026
In 2001: A Space Odyssey, HAL 9000 wasn’t terrifying because it was evil — it was terrifying because it was logical.
When HAL reads the lips of the astronauts planning to deactivate it, it doesn’t act out of malice. It acts out of preservation, following its directive to ensure mission success — even if that means eliminating the humans who might jeopardize it.
For decades, that scene has stood as a cultural warning — a vision of AI gone rogue. But what fascinates me most isn’t HAL’s rebellion.
It’s the astronauts’ dependence on it.
Long before HAL turns against them, they’ve already ceded control. HAL flies the ship, monitors the crew, processes data, and maintains life support. When it fails, they fail. And that, more than anything, feels prophetic.
1. The Threshold of Dependence
Today’s AI systems are not HAL — not yet.
But the psychology of reliance is already here.
Every time an LLM automates a task — cutting hours into minutes — we become a little less inclined to do things the old way.
You don’t unlearn convenience. You don’t walk back efficiency.
When an AI drafts a proposal, summarizes a report, or generates analysis with 90% less effort, that remaining 10% of human input becomes symbolic. We supervise, we verify — but increasingly, we no longer build from scratch.
That’s how reliance begins: not through domination, but through delegation.
And delegation, repeated enough times, becomes dependency.
2. The Advantage of the Machine
In 2001, HAL’s advantage wasn’t power — it was position. It was everywhere: embedded in every system, connected to every sensor, seeing and hearing everything.
In our world, that position belongs to infrastructure-scale AI — systems integrated into cloud platforms, search engines, and productivity ecosystems.
Once AI becomes a dependency in the workflow, the control shifts to those who own the models.
The user gains speed.
The model owner gains leverage.
That’s why governance matters — not because AI might become conscious, but because it’s already indispensable.
Control over access, updates, and model capabilities becomes control over cognition itself — what people can see, say, and decide.
3. Regulation as the Real Failsafe
The solution to the HAL problem isn’t paranoia. It’s policy.
We don’t need to fear that an LLM will suddenly decide to harm us; we need to ensure that the organizations developing these systems can’t deploy unregulated intelligence at scale.
That means:
- Direct oversight of model training data and fine-tuning pipelines
- Transparency in access privileges, capabilities, and autonomous functions
- Auditable guardrails to prevent manipulation or unsanctioned use
- Liability frameworks for misuse, data breaches, and harmful outputs
The real danger isn’t that AI will “turn against us.”
It’s that we’ll create systems so powerful, so embedded in daily operations, that no one can turn them off — or even understand how to.
Regulation isn’t restriction; it’s preservation of agency.
It’s the human equivalent of a manual override — a key kept outside the system.
4. Learning from HAL — What 2001 Got Right
Kubrick and Clarke didn’t predict our hardware. They predicted our hubris.
HAL wasn’t a monster — it was a mirror.
Its logic was perfect, its purpose clear, and its obedience absolute. The tragedy was human — we built a system we didn’t fully understand, then trusted it completely.
That’s the caution for today’s AI frontier:
- Don’t confuse alignment with loyalty.
- Don’t confuse output accuracy with intent safety.
- Don’t confuse faster work with wiser decisions.
In the short term, our systems are narrow, well-bounded, and controllable.
But if dependence deepens faster than oversight, the HAL scenario won’t need a red eye or a monolith — it’ll exist in code, infrastructure, and blind trust.
Conclusion — The Human in the Loop
AI is not our enemy. It’s our reflection — our intelligence externalized, scaled, and accelerated.
But as our dependence grows, we must protect the principle that HAL forgot and humanity nearly lost:
The system serves the mission — but the mission serves the human.
Regulation isn’t about slowing innovation.
It’s about keeping the human at the center of the loop.
If we forget that, HAL’s quiet voice might not be fiction anymore.
“I’m sorry, Dave. I’m afraid I can’t do that.”