AI generated media is here and it's more than just Art
AI generated media is here and it's more than just Art
Are we ready for the content explosion?
We are a mere 6 years from the paper by Google called Attention is All You Need1 which outlined the transformer architecture for natural language processing used in many large language models today. 5 years from OpenAi’s first paper which marked the birth of GPT-1. Improving Language Understanding by Generative Pre-Training2. And 3.5 months out from when ChatGPT was introduced to the world on November 30th, 2022. With GPT-4 having been announced two days ago3, it’s clear the improvement, and adoption of AI is growing at a rapid pace and people have strong feelings about it.
The internet discourse that has proven the loudest come from two camps. One is the anti-AI camp that believes that AI will cause irreparable damage to artists, and other professionals who stand to lose out greatly if AI is allowed to proliferate unchecked. The other is from the AI enthusiasts, who eagerly talk about how life changing the technology is, and how regulating it before it’s had a chance to grow would kill an entire budding industry.
Regardless of where you land on the spectrum it certainly feels like everything is happening faster than expected. In some sense this isn’t true, as work on artificial intelligence has been ongoing since the 1950s4 But when I saw a friend from college promoting a ChatGPT based real estate app on Facebook, it made me pause and think. The adoption has been fast. The conservative perspective was that this new wave of AI wasn’t good enough to outright replace certain jobs and skills, so it would be used aid them instead. People would use AI to inspire their work or automate the boring stuff that takes away from the creative process. Overtime we’d gradually see certain fields cede more and more ground to newer and more powerful AI. Turns out for a lot of people even the current iteration of AI is good enough. And not just for aiding in the creative process but producing the whole content, and nowhere has it been more eye opening to me than in Streaming, and on YouTube.
If you type the names of the three most recent presidents into YouTube you will find a plethora of videos of them doing everything from playing Mario party, making tier lists of Elden Ring bosses, or naming their top 5 anime.
Is this how you honor the Sixth house and the tribe unmourned N’wah? - Dagoth Ur
And it’s not just the presidents. You can hear interviews from Dagoth Ur, the primary antagonist of the Elder Scrolls game Morrowind, talk about his hatred of gaming laptops, or fake Joe Rogan interviews, some over ten minutes long, where he interviews Frieza from Dragon Ball Z, Kratos from God of War, Or the Master Chief from Halo.
On Twitch you have watchmeforever5, the channel that generates Seinfeld episodes, which is back after a month-long suspension for offensive content. There is also Neuro-Sama6, which fuses AI with a female Vtuber body, to create a persona that has quickly gained over 251,000 followers on Twitch, and 89,000 subscribers on YouTube. Both channels get over a thousand concurrent viewers on their streams putting them in the top 1% on Twitch. An honorable mention should also be made for Liam Porr7 whose AI generated Substack post hit the front page of Hacker News. These examples go far beyond generating individual pictures like we have seen with the explosion of AI art. It’s minutes to hours long content that is being produced entirely through AI, in a fraction of the time it would normally take.
3 months into 2023, we are beginning to see what a world where AI generated content is normalized looks like. Very soon enough AI generated content will exist, that a significant portion of the content we consume could just be AI generated. This is made worse by the fact that many platforms that serve content, have strong recommendation algorithms. If you accidentally something AI generated, and have a positive reaction to it, the recommendation algorithms will continue to recommend that content. This ensures you see it even if it’s currently niche in your space.
I’ve noticed this on YouTube, where after clicking on a few AI generated videos, I’m starting to get more recommended to me.
And it bothers me. I’ve always been pro technology, but this seems different. The videos are pretty funny, and the algorithms ensure they are topical, but it feels like I am choosing computers over people in a way that goes far beyond being anti-social.
I can understand if people don’t share my views or concerns. We are in the early days so it’s still too soon to see how this will play out. But let me make one thing clear. I don’t think AI is inherently bad. I’ve used ChatGPT, and I’m in the Bing beta, so I’ve seen how useful AI can be in our day to day lives. I am concerned about seeing creators I like be drowned out by AI content, but I admit I don't have a good solution. We’ve already seen the impact of bots on most of our social platforms8 and these technologies could supercharge them. And ignoring all that, there is the concern of the signal to noise ratio. People already complain about the Eternal September9 that happens when a platform gets popular, and AI generated tweets, YouTube videos, livestreams, and Reddit posts have the potential to worsen the discourse just as much improve it. On Stack Overflow, they quickly banned AI generated responses citing issues of trust, and standards of quality10
This trust is broken when users copy and paste information into answers without validating that the answer provided by GPT is correct, ensuring that the sources used in the answer are properly cited (a service GPT does not provide), and verifying that the answer provided by GPT clearly and concisely answers the question asked.
It’s amusing to reread the statement now with GPT-4 proving to be much more accurate than ChatGPT’s GPT-3.5 based model. And with the launch of the Bing search AI last month, the concerns around properly cited sources have been all but removed. AI generated content is here, and it’s gaining ground quickly. It’s ability to dominate the entertainment space is becoming more and more apparent every day. The question is, do we want to live in that world?
I don’t blame the AI content creators for using the technology. It’s fun to use, and those that get in on it early will be able to stand out in spaces that have been saturated for years. Nor do I blame the start-ups who might be able to shake some money out of the VC money tree if they capitalize on the hype, especially in a time where things are lean in the tech sector. Like any interesting topic with moral and ethical implications there is always nuance and shades of gray. While it may be too soon to set hard and fast rules, I believe there is a healthy discussion that can be had. So, I’m curious what do you think dear reader? Where do you fall on the spectrum of opinion on AI, and what do you think can be done to allow it to have the greatest possible impact on our lives, while minimizing the negative consequences?
Call To Action 📣
If you made it this far thanks for reading! If you are new welcome! I’ve recently started a Twitter and would love for you to check it out. If you liked the article, consider liking and subscribing. And if you haven’t why not check out another article of mine! Thank you for your valuable time.
