The AI Daily Brief: Artificial Intelligence News and Analysis - Why Are People So Scared of Sora?

Episode Date: February 19, 2024

NLW discusses the rise in anti-AI sentiment that seems to have followed the public announcement of OpenAI's new video generation model Sora. He provides a set of explanations for which this particular acrimony is arising now. INTERESTED IN THE AI EDUCATION BETA? Learn more and sign up https://bit.ly/aibeta ABOUT THE AI BREAKDOWN The AI Breakdown helps you understand the most important news and discussions in AI.  Subscribe to The AI Breakdown newsletter: https://theaibreakdown.beehiiv.com/subscribe Subscribe to The AI Breakdown on YouTube: https://www.youtube.com/@TheAIBreakdown Join the community: bit.ly/aibreakdown Learn more: http://breakdown.network/

Today on the AI Breakdown, we're asking why people are so scared of SORA. The AI Breakdown is a daily podcast and video about the most important news and discussions in AI. Go to Breakdown.net network for more information about our Discord, our YouTube, and our newsletter. Hello, friends, quick note before we dive into today's episode, since it is technically a U.S. holiday, I'm doing something a little bit different. Just a main episode focused on something that I've been observing all weekend. We'll be back tomorrow with our normal brief followed by main format. But the other thing that I wanted to quickly note is that I have just opened up registrations

again for the March version of the AI breakdown education beta. Now, since we started this in December, a lot has evolved. The goal is to create the single biggest, most useful repository of tutorials, case studies, and challenges anywhere to help you learn how to use AI tools, and we're up to over 100 pieces of content in that library. Throughout March, we'll be adding dozens of new pieces each week, sharing it with you via email as well as a private Discord just for community members. but I'm also excited to announce that we'll be moving our beta testers to a dedicated platform that we've been working on.

That platform will allow you to browse all the lessons, save for later those that you want to catch up on, and much, much more. To learn more, go to bit.ly slash AI beta. That is bit.l.ly slash AI beta. Registration is open from now until next Monday, February 26th at 1159 p.m. Eastern Time. And now let's get to the show. Welcome back to the AI breakdown. Last week, one of the big announcements alongside Gemini Pro's massive million-plus token context window was, of course, OpenAI's new video generation model, SORA. For many in the AI space, SORAL was an incredibly exciting moment.

In much the same way that last year we saw incredible advances in image generation, Sorrel really makes it feel for the first time, like that sort of capacity is coming to video as well. Almost immediately, AI Dreamers started coming up with new ways to use it. 11 Labs even created a way to automatically add audio to videos to make them a complete dynamic experience. Meanwhile, over in the non-tech part of the world, the reaction was very different. Over the weekend, the AI Safety Means account tweeted, poll after poll shows Americans as rabidly anti-AI, but my friends are techno-optimists so it didn't feel real. But since Sora, holy shit!

Look at these tweets. Mark my words, within five years, if we're still alive, there will be 100,000-plus person protests. Oh, and these protesters won't be rationalist nerd. It'll be actual degrowthers, unions, et cetera. And they won't advocate for nuanced things like pausing AGI, they'll demand a full stop. The current AI discourse is mostly a few techno-optimism factions arguing, but this will be a historical footnote.

EACs who spit in the face of even voluntary self-regulation from AI orgs are just asking for heavy-handed over-regulation. So let's take a look at some of the tweets that he was referring to. On a post of the litter of golden retriever puppies playing in the snow, Midas tweets, I need this shit to be illegal now, getting 205,000 likes and 11.1 million views. Charlie Moist Critical says, I'm struggling to think of a single positive thing

making realistic AI-generated videos like this will bring. It's all just net negative and dystopian. 8.2 million views, 155,000 likes. Megan Rose Ruiz writes, Gen AI is an effing insult to humanity. Image and video are no longer historical documents. With AI, images will become nothing more than our entire visual history shoved into a meat grinder

and served to us meaninglessly for profit. A craft I've dedicated my entire life is being replaced by an image predictor slot machine. 1.9 million views, 71,000 likes. Horizon writes, okay, but is there genuinely any benefit to this technology existing? Takes away jobs, old people and a hell of a lot of younger folks will be easily scammed, more garbage AI generated social media content, devalues human art. I can't see any positive to this at all.

41,000 likes. Anisa Senussi writes, AI should be doing all the useless corporate jobs nobody wants to do, not the creative work humans have been crafting since the dawn of intelligent thought. Come on now. 444,000 views, 33,000 likes. You get the picture. And what was interesting is that this was not just an account that focuses on AI risk going

out in cherry-picking examples. This is something that I noticed a massive uptick and negative sentiment around as well. Particularly in other social media channels like TikTok, over the past weekend, I've just seen endless amounts of really, really negative responses. So what is it about Sora that is driving such fears? What one part of it, I think, is the relative influence and importance of video as compared to other mediums. Video has of course become the dominant channel for basically all discourse at this point. Between Instagram Reels, YouTube shorts, and TikTok, everyone is a mass consumer of video,

and many are video creators as well. So I think one reason that Sora is generating such a stronger response is just the proportional importance of video. But the second piece of this that I think is really interesting and different than what I've seen before, is that Sora is producing specific fears around specific issues. Up until now, the negative A-I-I-I-I-Rexecutive. discourse has been largely around two things, either on the one end of the spectrum, artists having their work stolen or devalued or copyright infringement, or on the other end of the spectrum, actual existential risk. In other words, the people who have been concerned about AI, at least publicly so, tended to fall into one of those categories with the content that they were sharing about it.

Now, of course, if you add media into the mix, a third topic, which is AI job displacement, comes up frequently, although interestingly it happens a lot more frequently in the media than it does, at least in my anecdotal experience, with individual viral content. The response around Sora, while there has been some amount of the artist thing, this one is really breaking out of that in a huge way. People are having specific fears around specific issues. For example, deep fake porn and sexual exploitation. Here's Ali Rooker from TikTok. I am losing my mind thinking about how many women's lives are going to be ruined over this AI video bullshit. Don't tell me I'm wrong. Don't tell me it's not going to be used for what we

all know it's going to be used for. Because it already happens, and now we're just going to add video to this. Nothing good that could be created with this justifies the evil that will be done with it. Like literally my first thought when I started seeing these videos was like, I need to delete my social media. Like this is not safe. This is not the only thing that this is going to be used for for evil. Like the political corruption. And this of course gets us to a second specific concern which is around political deep fakes. Here's Megan B. Rice from TikTok. I cannot stress enough how dangerous the SORA AI thing is, because people are looking at it through the lens of there goes human expression,

there goes art and creativity, but military, governments, media companies would also have access to this. We would not begin to even know what was the truth. Governments and military could completely change the narrative. The oppressed can easily. be changed to the oppressors. The oppressors can easily be flipped to the oppressed, and it would look 100% realistic. We would not know who to fight. Some people are getting even more specific. Some people are even concerned about being accused of fake crimes.

There's a chance we might start getting charged with crimes we never even committed. So I'm assuming you all saw those SORA AI videos. If you haven't, it's very, very realistic footage completely AI generated from a text prompt. These videos have people very, very concerned for the future, specifically when it comes to criminal justice. Many people are concerned with the possibility that AI generated videos will be used in court in order to pin people with crimes that they never committed. I mean, if you just take a look at SORA AI now and imagine what it'll be in just a couple of years down the road, I think that being charged with a crime because of an AI video is almost guaranteed to happen in the future. Okay, so we've talked so far about the relative importance of video in society

and about people having specific fears about specific issues this time. But there's also another thing going on, which was hinted at in that last video as well, which is that it feels to me like people are finally getting a sense of the true speed at which AI is evolving. Just 10 months ago, people were meming and making fun of videos of Will Smith eating pasta that were AI video generations. And now here we are with this incredibly advanced, nearly indistinguishable AI video, there is a raw sense of power that is being felt, which perhaps hasn't been in the future because we haven't had such a clear comparison while people were paying attention.

Now, there is another set of issues, though, that have to do more with technology and society than just AI on its own. It used to be for a very long time. The technology was basically a priori assumed to be good for us. New technology, with the exception perhaps of war-related technology or nuclear power was just always a net value to society, or at least in people's perception it was. Social media officially broke that pattern. There are tons of reasons for this. It didn't help that the first dominant social media platform in Facebook had so much ick involved with it that

we all learned about because of the social network movie and everything else surrounding Facebook. But when you think back, I think that it's arguable that the last time that many people had a net positive view of social media. In other words, a sense that it was a sense that it was was a net good for humanity versus a net bad, was the Arab Spring protests in 2011 and 2012. And indeed, for many, this was what the promise of the internet was supposed to be. A 26-year autocratic leader like Hosni Mubarak being forced out of power in a mass popular protest enabled in large part by the coordination capacity of social media. Since then, however, the forces pushing us to feel that social media is net bad have gotten

louder and louder and louder. Of course, we had the 2016 elections, where Democrats in the U.S. accused Facebook of enabling Russian bots to undermine democracy and allow Donald Trump to beat Hillary Clinton. And if that was the political left's cross to bear, on the political right the next few years, saw this never-ending theme of accusations of de-platforming, culminating, of course, in the actual suspension of Donald Trump's Twitter account. So basically, you have both sides of the political mainstream in America telling their people that social media sucks, even if for totally different reasons. But even if there hadn't been this political response, there's just generally been a feeling

that many, many people have, that social media just isn't all that good for them, and it's probably not all that good for us. It is unbelievably easy to find statistics like this one, that 92% of parents think social media and the internet is having a negative impact on their kids' mental health. And then, of course, now there's a geopolitical dimension, with TikTok constantly on the verge of being banned, as effectively a mass data collection and disinformation tool for the CCP. And so the point of all of this is that resentment around technology has been building for more than a decade. A sense that technology is not a priority good, in fact that we should be more skeptical of it has been building for more than a decade. And now, that latent force, that latent

skepticism that has been growing, is combining with a totally different phenomenon, which is AI. Now, people can hem and haw as much as they want, but it is unbelievably clear to anyone paying even a little bit of attention that artificial intelligence is going to have a dramatic impact on the economy, on how we work, on how we learn. You can't open up the internet without stumbling across some new statistic about just how impactful it's likely to be. One recent one that got a lot of attention around Davos earlier this year was the IMF suggesting that in developed countries like the US, AI would impact 60% of jobs. So you have this latent mistrust of technology, plus the incredible likely impact of AI. And then people are looking around

and seeing that it's this tiny handful of companies making decisions about these things that will have or seem like they will have such a dramatic impact on their lives. Now, I do have to say one other thing here, and this is an important caveat. Whether you think TikTok is actually a tool for the Chinese government or not, our opinions are absolutely weaponized by algorithms right now. After going back on my TikTok feed and grabbing those videos that I shared with you earlier, the next hour of content was almost entirely super-inflammatory left-leaning content. Endless TikToks about how there's no way to be a billionaire ethically

and take your pick of issues that might appeal to that specific type of person. my whole feat was that. No more of the seasonal decorations, magic cards, and guitar stuff that I normally have. And so, of course, we have to remember that whatever latent feelings of frustration, skepticism, or fear are there, they are being massively amplified by the networks that are at least partially responsible for some of those fears in the first place. Now, one of my concerns is that at least when it comes to the loudest voices, the lines are drawn so aggressively right now that it doesn't leave a lot of space for meaningful discussion and discourse. For example, on my tweet about seeing a big uptick in negative sentiment around AI and Sora specifically, Mr. Schroom, who calls

himself a technology brother focused on the positive economic impacts of artificial intelligence rights, I've noticed that a lot of people that are pro-AI have lost patience with these anti-AI people, and it results in a lot of name-calling in more toxic environment. I get it, but I'm really trying to be as civil as possible in assuming good faith and point out the nuance of how I believe the positives of new technology generally outweighs the negatives. To that, fierce anti-AI advocate Jeffrey Miller responded, a lot of us anti-AI people have lost even more patience with the pro-AI people. The battle lines are drawn. When I asked genuine question, do you think there's space in the middle to reset the debate? I got no response. Now, obviously given how much time I spend in AI, not only on this podcast,

but also producing multiple tutorials, lessons, case studies every day for this AI education beta, hoping to teach people about how to use AI, I think that there is immense opportunity for good. I am empirically more optimistic than pessimistic. I think that fears are being exacerbated for value in both intentional and unintentional algorithmic ways. But I also think that society should have the opportunity to decide whether something that's being built serves them or not. I think that if you took the vast majority of people who weren't already radicalized in one direction or the other, there would be an immense open space for meaningful discussion about what we really wanted out of AI and how to drive it in a direction that achieved that. My own little piece of this increasingly is

trying to give people the tools to at least leverage this for their own lives and careers so that they feel agency and a shift which is coming whether they would have chosen it or not, but can maybe actually take advantage to create the future they dreamed of. Whatever the case, I think that SORA shows us that this conversation is going to get more dramatic, not less so, and that is just beginning. It's certainly one that I'll be paying attention to, and you know I will share it here when I find things worth sharing. For now, that is going to do it for today's AI breakdown. Until next time, peace.