The Potential Dark Side Of Advanced Artificial Intelligence & Machine Learning Technology
The Dangers of Deepfakes
I received some positive feedback about a post from last week that highlighted an interesting application of artificial intelligence.
The Replika story is an example of what could be considered a positive application of technological advances. However, like with most new and innovations, people will often find a way to exploit it and use it for different purposes.
Today, I want to discuss how one particular type of technology known as deepfakes has emerged to change what we think we know about the truth.
Deepfakes are defined as synthetic media that use artificial intelligence and machine learning to generate altered and simulated videos and images that misrepresent what really happened.
Simply put, it’s a manipulated video that looks extremely real but isn’t. Think about it as high-quality Photoshop editing to both the audio and visual of a video.
You’ve probably seen a variation of this, likely through social media. This list of examples illustrates some of the more popular deepfakes that have plagued the internet.
The problem with depictions like these is that they can create a foundation of fake news, fraud, and in some cases, hoaxes and conspiracy theories.
The most extreme example of this would be if a false video was constructed of a powerful figure (for example, the President) saying or doing something that they didn’t actually say or do. That could have unprecedented implications.
This isn’t entirely unrealistic either.
A more basic and more likely use-case of deep fakes would be to perform practical jokes or fun social media posts by using tools like Reface or FaceApp.
Because the lines on the moral implications of this technology are still so blurry, a few large tech companies have created initiatives and programs that make it easier to spot and label these manipulated forms of media. The availability of these labeling features isn’t widespread yet, however.
Still, some in the industry have argued that the benefits of deepfake technology could outweigh the drawbacks.
For example, it’s been cited that this could be used to develop movies, TV, and film productions much faster and cheaper since actors wouldn’t have to film every scene as they previously have been forced to do.
Also, e-commerce applications could be used to show consumers how customized products or clothes can look on various body types, etc. The list goes on.
I think this debate will continue to be an important one, particularly when (not if) we end up seeing a large scale use of this technology either by a popular brand, political party, or media outlet.
Where do you think we should draw the line on this? Should deepfakes be allowed?