OpenAI's Sora has left AI experts either enthused or skeptical. It's left most everyone else terrified – Yahoo Canada Finance

Hello and welcome to Eye on AI.

It’s been a week since OpenAI unveiled Sora, its new text-to-video generative AI model it says can turn short text prompts into strikingly realistic videos up to a minute long. The videos shared thus far have been received as thoroughly impressive and a giant leap for AI video generation (said with some reservations to account for the fact that OpenAI hasn’t demonstrated the model actually working or released a technical report). But today, I’m diving into a different through line of the reaction to Sora, and that’s fear.

Outside the tech and AI communities’ continued debate about accelerationism versus doomerism, there are a whole lot of everyday people who are personally terrified seeing this technology progress so rapidly before their eyes. This is not exactly a new response to AI, but it’s one I’ve seen growing and that seemed to hit a new level over the past week with Sora. While scrolling through my For You Page on TikTok, for example, I saw video after video of everyday users who do not typically post about tech news expressing their fears about Sora and AI. One account I follow that typically posts about pop culture and pottery took a break from sharing techniques for making succulent planters to lament about the announcement, asking “Is anyone else concerned that AI is going to be the downfall of society or is it just me?”

“RIP reality,” replied one user.

“I give humans 5 more years tops,” commented another.

“I’m genuinely scared to death,” wrote another user, along with countless others expressing the same sentiment.

It’s easy to see why. We know this technology will be used to deceive, create harmful deepfakes, generate and spread disinformation, and sow chaos. It’s already happening—and at a critical time with democracies on the line. The entire world has already been reeling from these issues as they’ve been perpetuated by social media platforms, and now generative AI is poised to add fuel to the fire. Perhaps even more importantly, technologies like ChatGPT, DALL-E, and now Sora are being positioned as more efficient writers, visual artists, and filmmakers, potentially swallowing up the creative arts that people enjoy, that empower us to express ourselves, and that makes us feel human. And what exactly will we all gain from this? Cheaper stock footage? Movies created by AI? The creators of AI will certainly gain unfathomable amounts of money and power, but for everyone else, it’s not exactly clear what the benefits are and if they’ll be worth the costs.

Story continues

The way tech companies are going about it isn’t helping either. The shared research community-oriented approach to building AI that existed for decades went fully out the window when OpenAI released ChatGPT and was quickly replaced with secretive development, rapid commercialization, shareholder demands, corporate lobbying, and pursuits for market dominance and $7 trillion valuations.

“At the top AI firms, employees are now being asked to keep their accomplishments secret, and in some cases, stop publishing research papers altogether, upsetting the symbiotic relationship between the pursuit of profits and scientific discovery,” wrote Reed Albergotti in Semafor this past week.

That’s not to say there weren’t commercial incentives before, and it doesn’t account for the passionate community of AI professionals working to bring transparency and accountability to the field. But many everyday people feel burned by the impact digital technologies have had on their lives and society, by tech companies themselves, and by the state of capitalism overall. They don’t feel technology has lived up to its promises, and they’re seeing tech companies rake in record amounts of wealth (not to mention lay off thousands) while they struggle to meet their basic needs. People are already wary tech companies will prioritize profits over their best interest, and so the increasingly secretive development, lobbying, and so on isn’t exactly furthering trust in AI or the companies creating it.

It’s important to point out that many new technologies were initially met with fear only to become accepted and even critical in our lives. The introduction of electricity sparked (valid) safety concerns and an entire anti-electricity movement. Photography, while welcomed by some, received massive backlash in the arts world as people viewed it as a cheap shortcut that would supersede the true art of painting (sound familiar?). And anyone reading this likely witnessed the advent of video games—and reactionary concerns they’d foster mass violence and addiction—only to see those fears haven’t exactly come true, research actually suggests positive social and cognitive impacts associated with gaming, and that it’s grown to become a massive $200+ billion industry—larger than the film and music industries combined. Even the modern mirror, a now everyday household item, was widely feared upon its introduction as it was thought to pose a moral threat to society by encouraging vanity. Many said the same about front-facing cameras and selfies.

Now that we have some historical context, we can talk about what feels different this time around, because AI does feel different. The TikToker I mentioned earlier put it well when he said he thinks there will be a day in the next few months when we wake up and cannot tell the difference between what’s fake and real online. In many ways, this is already happening and people cannot tell the best AI-generated images from real photographs. Electricity is in fact dangerous when not properly set up and managed. And while mirrors and photography were unknown, they at least reflected reality. Generative AI, on the other hand, distorts and deceives us of reality by design.

Over these past several months, I’ve seen some argue these generative AI tools are not enabling people to do anything they couldn’t already do in Photoshop or After Effects. Photoshop is probably the closest proxy for software that can distort reality, and we could debate the positive and negative impact it’s had. But I think it’s becoming clearer every day that these AI tools are far more powerful. More importantly, they manipulate reality via a black box, not a human controlling a Lasso tool. And critically, Photoshop takes serious time and skill to learn (and is of course now offering generative AI tools), whereas anyone who can write a sentence can prompt AI tools to create whatever they want almost instantaneously.

As someone who’s covered AI for almost a decade, I’ve gone from few people around me even knowing what AI is to overhearing a table of teachers out to lunch desperately trying to figure out how to handle it in their classrooms, receiving concerned messages from family members asking me to explain it, and encountering fear of AI from everyday people with every swipe on social media.

Sora may not be generally available right now, but just the knowledge of its existence is already spreading a ripple of uneasiness through society. Well, at least we can all console ourselves with the thought that, as the AI accelerationists like to say, the AI-generated video you’re watching right now is the worst AI-generated video you’ll ever watch from here on out. Wait, why don’t you look happy? 

And with that, here’s more AI news.

Sage Lazzaro
sage.lazzaro@consultant.fortune.com
sagelazzaro.com

This story was originally featured on Fortune.com

Adblock test (Why?)

Source link