Entering the Age of AI.
What could the future bring?
We are at the birthing point of influential AI.
Motorised vehicles changed the transport industry. Telephones changed the communication industry. The internet changed information consumption.
AI? AI changes the world.
I would tell you to brace yourself, but for what exactly? The future has more unlimited potential now than at any other point of time in history. We are either standing at the edge of a steep precipice or of something great, it’s exciting, it’s scary and it’s completely unpredictable.
We are about to enter the age of AI.
ChatGPT
I was going to describe what ChatGPT is, but I thought it’d be better just to ask it itself.
Ask it any question, and it will spit out a coherent, well-thought-out answer in seconds - if you can call it a ‘thought’.
I used a similar program to help me write a script for my video on mindfulness. I gave it a few prompts and it wrote out everything I needed in no time at all. It shot goosebumps down my spine. I saw the future unfolding right in front of me.
When AI further improves and becomes a ubiquitous tool, is there much point in continuing this hobby of mine? Articles can be written with AI assistance, kids are using it to write their essays, who’s to say it won’t be able to soon write entire novels?
Creativity was always thought to be one of the last avenues AI would take over but it turns out it was one of the first.
Artists
It looks like even the creative domain will be taken over eventually.
Let’s try to peer into the future of entertainment a bit more: AI can create still images with human assistance, but surely it can’t create entire movies and TV shows?
You’d be wrong. An AI could analyse hundreds of great movies, find patterns that make them appealing and recreate its own full-length movies. It would need some human guidance and overlook, but it would be as simple as giving it a few prompts and then a few hours to render. If something doesn’t look right, the human can use a tool on the AI program to highlight it and changes will be made within a few minutes. This can then be done for every scene or part of the shot that needs to be adjusted. We could soon go from an industry where it took thousands of people thousands of collective hours, to one person with a powerful computer and an AI program.
The entire future of everything will change, and it’s probably closer than you’d think.
Warfare
What would you do if you saw secret recordings of your country’s leader talking about how they were planning to tank the country’s economy for their own benefit? Well, you’d probably demand an internal investigation at the least, possibly even an immediate resignation. What if the government then came out saying it was a deepfake video? How can you, as an individual citizen, verify and determine the truth of the matter? It would be expected the government would deny this video, but we should soon expect such videos to be created to ruin the reputations of people. Hostile countries could create deepfakes in order to sow discontent and distrust within the population. Nations are already using social media for this purpose, but deepfakes will exacerbate the effects of their techniques.
In other words: deepfakes will become weapons of war.
Intelligence agencies can (in fact I predict they will) use deepfakes to their advantage. Wars and coups are won when the minds of the general public are successfully changed. Deepfake videos of enemy leaders advising citizens or troops could be disseminated across the frontlines, wreaking havoc and confusion. You’d have people obeying what the deepfake told them to do, and others trying to confirm if it’s real. The leader would have to put out a video saying it’s false, but… How would people know which one is the deepfake?
I hope you’re starting to see the problem here.
At this point in time, most of us are able to spot deepfakes. There’s usually something a little bit off and robotic about them. But we’ve come a long way in recent years. Give it a few more years, even a few months as the creator of many recent deepfakes says:
What do we do?
Here’s ChatGPT’s recommendation:
Let’s go through each of these points.
Points 1 and 5 are essentially the same: we need to increase our scepticism of every single video we see. Remember, fake news travels around six times faster than real news. So if you see a video that makes you want to pull your hair out in frustration or confirms what you already believe, chances are it’s not real.
Point 2: when deepfakes become indistinguishable, this is no longer a viable option.
Point 3: if you’ve been alive for the last few years, you’ll know ‘trusted news sources’ and ‘fact-checking websites’ can be compromised and constantly get things wrong.
Point 4: this creates a way to see a video’s digital fingerprint. In the same way you can use reverse image search to see if an image is original or not, we might be able to do the same for these videos.
Point 6: this is certainly useful, however, an arms race will be created between deepfakes and deepfake detectors. We will again be stuck in limbo not knowing which one to believe.
All the above methods can be bypassed. If an intelligence agency - foreign or domestic - wants a video to be widely disseminated and believed by a portion of the general public, it will happen.
I can’t think of a way to ensure we will never be fooled by deepfakes. This could end up being a serious problem, such a serious one that it tears apart our social fabric. If it’s divisiveness they want, divisiveness they will get. Deepfakes are the perfect tool for this job.
I asked one of the most interesting, AI-focussed people on Twitter what he thought was a solution to the upcoming problem of deepfakes could be:
The future of truth is not looking bright.