So far, not that frightened. The technology is still in its infancy, and has only been used in what researchers call “image-to-image translation,” or altering video clips and photos in small ways such as changing the setting from night to day, changing human subjects’ hair color, or switching a dog to one of another breed. Still, that’s pretty impressive if you think about it. Nvidia’s Ming-Yu Liu says that their system is the first to be able to do so simply by ‘imagining’ the new image or scene, as opposed prior similar systems which faced the problem of having to compile massive sets of data based on prior examples and extrapolating from those data:
We are among the first to tackle the problem, [and] there are many applications. For example, it rarely rains in California, but we’d like our self-driving cars to operate properly when it rains. We can use our method to translate sunny California driving sequences to rainy ones to train our self-driving cars.
The potential applications of this technology have lead some to wonder if similar AI networks might mean the “end of reality as we know it.” Whatever that means. Unless the whole universe spontaneously blinks out of existence, reality’s not going to end anytime soon. But I see what they mean; if completely real-looking video and audio can be generated by these AI networks, and those could be fed into, say, an advanced augmented reality setup or even some of the more Matrix-like brain-computer interfaces being developed, we could soon see the lines between virtual reality and physical reality start to get more difficult to distinguish. Until you take your headset off, that is. But what about if when these experiences can be transmitted directly into your brain’s sensory centers?
While it’s unlikely we’ll see the end of any reality, we might see the creation of fully-fledged alternate realities. Without a doubt, this technology will someday be used to distort or obfuscate the truth here in our own reality. What is reality other than what we make of it, anyway? Recent events have shown us that it’s getting more and more difficult to discern truth from fiction in mass media; what will happen once completely real-looking video can be conjured up from the twisted imaginations of rogue AI systems? Of course, the same fears arose over the invention of moving pictures. Is this just the latest advancement in graphic and animation software, or could something more nefarious be brewing?