When AI makes mistakes, people must think about how to fix those errors.

 When AI makes mistakes, people must think about how to fix those errors.

We all hear stories about things, like cases, where the AI thinks that things like commercials are traffic signs like speed limit signs. Then that AI decides to make the maneuver where it increases its speed. There are a couple of easy things like using the GPS where the maximum allowed speed is put in the system. That thing makes it possible to avoid those errors. And another way to make AI-based driving systems safer is to use interactive traffic signs. In that model, the traffic sign will equip with a system that allows it to tell what it is to cars. 

The system can base the secondary radar respondent technology. Those kinds of systems can be optical or (and) radio-based. The AI aims the laser ray or radio wave at a thing that it suspects is a traffic sign. When the AI sees traffic signs it aims thin radio signals to the traffic sign. And then the radio or optical system in that traffic sign answers the AI what it is. 

In the optical system, the traffic sign sends laser rays to the camera that the AI uses. Then the AI will not make mistakes with those kinds of systems. The interactive traffic marks will not be expensive. The simpler model is that the traffic signs are equipped with QR codes that are visible in UV or IR frequencies. That thing makes AI-driven systems safer. 

Interactive AI-based technology is complicated. That means the system needs very well test and development work so that it works as it should. The thing is that when people are starting to use AI or autopilot-driven cars they must read instructions. The AI doesn't mean that the driver can read newspapers and sleep while the car is moving.