Impact of Visual Recognition on Society

Greg Golden
3 min readNov 10, 2020

--

Image from TalkWalker

IBM’s visual recognition software has made leaps and bounds in the recent years, but it also has become more and more of a privacy concern when applied to real life. I was using their visual recognition software, and I realized that the software once trained can bring up privacy concerns. I saw this firsthand when I was trying to find Waldo using the service, and it was surprising how well it already was finding Waldo. It made me realize if I had trained this even more, then imagine the accuracy Watson would have with detecting Waldo.

Image from Financial Times

There are of course some concerns that arise with this sort of software, that inherently would not be bad, but as time progressed could make it worse. I always connect it to the data itself not being the problem, but who owns the data that is created. What are they going to do with my data? Are we going to end up in a Big Brother scenario? Will we start to see an erosion on our freedoms, because of it? These are just some of the many questions that can arise when we talk about visual recognition software. Do not worry though, it is not all bad when it comes to visual recognition — it can do many good things as well!

Image from IBM

I have been taking this class called Impact of AI on Society and have been gaining a better understanding of how AI services like visual recognition, and discovery services Watson provides. This class has allowed me to really see an in depth look of how these services work and understand how these services will affect the world around us. Visual recognition and other AI services provide a huge convenience for data collection and analysis. Visual recognition has helped us compare scans in the medical world and be able to pinpoint which ones may have cancer against those that do not. This is just one of the many advances in the world of visual recognition, that we never would be able to do without it.

It is not that visual recognition would be inherently a bad idea; it comes down to where this data ends up and who it ends up with. Society needs to keep this in mind as these services improve further and become more and more invasive in our daily life. I read an article about how we cannot stop the advancement of prediction machines, but we can better collaborate those machines with the human mind to innovate further. Our society needs to figure out how visual recognition is going to impact our lives and probably put a lot of people out of work. If we are able to do that, then our limits are endless and we won’t have to worry about these privacy concerns.

--

--

No responses yet