If you were looking forward to the day when a robot could be your trusted companion or reliable sidekick, you may want to think twice.

New research indicates robots can learn to be biased — that includes racial and gender.

Research led by three major universities found that a robot programmed with a widely-used internet-based, artificial intelligence system was more likely to choose men over women and White people over people of color. It also made assumptions about a person’s occupation after glancing at people’s faces.

The research indicated the robot learned to be prejudiced based on a popular model of AI that is being currently used on some robots.

The researchers were scheduled to unveil the findings in their report last week.

More about racist robots of the future can be found here.

Opinion: This all sounds Orwellian but calling these robots’ programming AI doesn’t seem fitting. If its racist and sexist judgments mirrored humans (those who programmed it to start with) then it’s hard to call it artificial… or intelligent for that matter.