Last month in Jakarta, I met an old friend. He picked me up in his new electric car. I knew it had advanced features and he mentioned that it was equipped with AI. I had seen similar technology on YouTube before, demonstrations of self-driving systems, automatic parking, futuristic dashboards. But this time I was not watching a video. I was sitting in the passenger seat.
At some point during the drive, he activated the assisted driving feature. The steering wheel adjusted on its own. The car kept a careful distance from vehicles ahead and changed lanes smoothly.
Later, it parked itself in a tight space with a level of calm precision I probably would not attempt in Jakarta traffic. It was impressive, not in a dramatic way, but in a quiet way that made me pay attention.
To be honest, it drove better than most human drivers I encounter on a daily basis. The system worked. It did exactly what it promised.
When AI Leaves the Screen
For years, AI has lived inside my laptop and phone. It writes suggestions, recommends content, summarizes information. Helpful, but contained.
Sitting in that car, I realized something had shifted. AI is no longer limited to screens. It is now steering physical objects, navigating public roads, and operating in shared spaces.
The moment technology moves from digital assistance to physical control, the stakes change.
The Question in the Back of My Mind
While the car was driving itself, one thought stayed with me.
What happens if this system is compromised?
Every software system is built by humans, and anything built by humans carries vulnerability. If a productivity app fails, it creates inconvenience. If a financial platform fails, it causes financial damage. If an intelligent vehicle fails, the consequences are physical.
The more we integrate AI into infrastructure such as cars, homes, healthcare, and logistics, the more digital risk becomes real-world risk.
How Trust Forms Quietly
What struck me most was not the technology, but how quickly I adjusted to it. Within minutes, I relaxed. I stopped watching the wheel so closely. I assumed the system would correct itself.
Trust does not form through philosophical debate. It forms through repetition. When something performs well consistently, we grant it authority. That shift often happens quietly.
We usually talk about AI risk as malfunction. But another possibility is misinformation. An AI system can present something confidently while being incorrect, and most people will accept it because the system appears precise.
In a search result, that may mislead us. In a vehicle, the consequences are heavier.
The problem is not capability. It is misplaced certainty.
Use It, But Do Not Surrender to It
I am not against AI. The drive was smooth and the technology is remarkable. It will likely improve safety, efficiency, and accessibility in ways we are only beginning to understand. The progress is real, and it deserves recognition.
But progress does not eliminate responsibility. The more capable the system becomes, the more important human judgment becomes. Convenience can slowly train us to disengage. And disengagement, over time, becomes dependence.
We can use AI and benefit from it. We should explore it, learn it, and integrate it into our work and daily life. But we should not surrender our awareness. Trust should be proportional, not absolute.
Because the moment we trust something completely, we stop questioning it. And the moment we stop questioning it, we stop supervising it.
Technology can assist us, but responsibility still belongs to us.