While reviewing articles regarding Autonomous Intelligence (AI) and how like any and all technology are able to be hacked. I feel new technology released each day and within the same day someone is trying and succeeding in creating a hack for malicious use.
The good news is, those reviewing for the bugs and items that can be hacked into it isn’t always the “bad” guys. A University of Washington student is studying AI self driving vehicles, how the computer systems see and interpenetrate road signs that have been modified in some way.
The modifications are not unlike what we see on a day to day basis, such as a stop sign with a sticker on it or graffiti. The research which is paper basis as stated this has not been tested on a self driving vehicle. However, the study shows with slight changes to a stop sign an AI self driving vehicle went through the stop sign. This was also shown with speed signs as well, the vehicle would speed up instead of slow down.
Though this has been tested, the proof is only for self driving cars which solely rely on video cameras and a computer to process the information seen. Cars such as Subaru’s with EyeSight which assist in braking and alerting the driver when a vehicle is in the blind spot.
Alicia's Blog 