AI and its (maybe) Horrifying Consequences

This post is an information hazard that could lead to the mental harm of yourself or the harm of others. Read at your own risk.

Where We Stand

By now it would be hard to believe you haven’t heard of ChatGPT or OpenAI in at least some capacity. The model GPT-3 took the internet by hurricane and sparked existential conversations about aspects we once considered exclusively human. There is rightfully a fear of AI taking jobs as Dall-E and GPT4 begin to encroach upon art and basic coding/webdev respectively. Right now we stand at the precipice of something great and truly terrifying. Ethical, moral, and practical debates are ongoing as we discuss AI’s place in our future … or our place in AI’s future.

The 'Arts'

Can AI Feel?

As humans there exists a drive in some of us to express ourselves in creative ways through the various arts. A painter can be moved by beautiful scenery and paint a picture to move millions, much the same way a musician draws on melodies to create music. But art doesn’t stop at music or paintings. Jokes, church sermons, bananas taped to walls, and even diagrams of submarines can be considered art. What constitutes art is bound more by how we view it rather than a rigid definition: It is at its core a way to feel, and therein lies the issue when we consider AI art.

In the world of overbearing corporations that do everything they can to save a penny there is no surprise that people who work industries are fearful of its potential for abuse. Not only as a means to replace artists, but to do so as an uncontestable cost. Admittedly the artists at immediate danger are those working with digital art. That poses a serious threat to any modern artist looking to utilize the readily accessible internet to create or propagate their art. Will their potential job be taken by one person mass producing art through prompts and an AI model? Let’s not even get started on the implication of proving that your digital art wasn’t made by AI and the hundreds of hours of footage new artists might need to ‘validate’ their claims.

Not Legal Advice

Feeling and art is overrated anyways. Onto a more pressing AI art topic: Copyright. AI models are created by training them on pre-existing images. If an AI model was created exclusively on copyrighted art as source material, what then would it be producing? Does the resulting image constitute theft if copyrighted materials were used to create the piece? Does any of this even matter if you can’t tell the difference?

To large corporations like those in the music industry: Yes. Unfortunately for the rest of us, good or bad, those giants who already control large amounts of copyright are going to be fighting tooth and nail against those who want to use AI for art. Up until they can abuse and use it for themselves.

The true terror of this AI music or voice technology is it seemingly has no purpose to serve the population without introducing a myriad of moral questions. Using 100 words to train an AI model with your grandmother’s voice so you can hear her again sounds sweet, but lets replace your grandmother with a late music artist XXXTentacion and you with Capitol Music Group. Now the creation of this voice model to produce more music would be anything but noble. It would at its essence be a selfish act to potentially taint the reputation of a beloved artist after his passing and make money in the process.

Catfishing Made Easy

And the icing on the cake is that this technology exists right now. The twitter post https://twitter.com/rpnickson/status/1639813074176679938 demonstrates training a Kanye voice model to sing lyrics that are fed to it. This technology doesn’t stop here though because as per DailyMail a Houston couple was scammed by a scammer copying their sons voice and asking for $15,000. Luckily the couple only could give them $5,000. What on Earth could this technology be used for besides malicious ends?

Dating apps are about to be destroyed from the inside out. One person catfishing can now create realistic chatting bots, realistic images, and realistic video chats containing audio of women and men that do not exist. Then once those models are created they can attempt to extort or scam money out of lonely people using popular dating apps. Now ask yourself this: is it even possible to ban an AI that is probably better at conversation than the average dating app user, and can artificially generate any verification you ask for?

Hackers Need Not Apply

Webdevs and Coders Beware

Let’s say you’re an entrepreneur with the desire to create a store to sell your services. Traditionally one would need to use a website building service such as Wix or pay a web developer to make a page the way you want. Now all it takes is some sentences describing what you want to let ChatGPT produce a functional website using standard coding practice. Even worse than that is someone with ChatGPT can offer their services as a web developer, use this AI to create websites, and sell their services for money they aren’t the ones earning.

Code generated from ChatGPT and other AI is unlikely to be susceptible to bugs and hacking. It uses proper documentation as source material which in most cases does not introduce vulnerabilities. This implies that both the coder and the penetration testers become obsolete and replaced by an online chatbot. This goes further though, need a specific type of functionality made with javascript? Ask the AI for it, implement the code, and feed it the errors you see if any. Debugging and coding has always been a highly marketable skill up until now. One AI chatbot user can now take the jobs of hundreds. Where does this end?

The End of the Road

This is not where we as hackers or tech enthusiasts can sit back and watch. This, is a call to action. The German hacker LiveOverflow on YouTube created a website to help combat the future of our uselessness: https://rAIsistance.com. In this site it eloquently describes ways to keep us relevant the only way us hackers know how. Say it with me, sabotage. We need to introduce bugs into code for us to fix. We need to rewrite decade old tutorials on secure code to introduce vulnerabilities that these new AI models will be adopting. We need to disturb trust between people before this AI can get off the ground and do it before us. Hackers are the scourge of society, not this trained technology. How can we survive without the demonization from mainstream media of our craft? Now get out there.