When Apple announced plans to build generative AI into the iPhone, it was practically official. The technology is now inevitable. Large-scale language models will soon be lurking in most of the world's smartphones, generating images and text in messaging and email apps. AI is already dominating web search, showing up in Google and Bing. OpenAI, an $80 billion startup that has partnered with Apple and Microsoft, feels ubiquitous; its ChatGPT and DALL-E auto-generation products are ubiquitous. And for a growing number of consumers, that's a problem.
Amid all this controversy and consumer anxiety, technology has rarely come to the forefront, or been forced to come to the forefront. To be sure, some Americans are excited about AI, but a recent survey found, for example, that a majority said they were worried that AI would increase unemployment. Another survey found that three in four people said they thought AI would be misused to interfere in the next presidential election. And many AI products have failed to impress. Google's “AI Overview” launch was a huge flop. The search giant's new bot cheerfully told users to put glue on pizza and that poisonous mushrooms are safe to eat. Meanwhile, OpenAI has been mired in scandal, infuriating former employees with a controversial non-disclosure agreement and allegedly scamming one of the world's most famous actors for its voice assistant product. So far, much of the resistance to widespread AI has come from watchdog groups, concerned citizens, and creators worried about their livelihoods. Now, consumer backlash against the technology is also starting to spread, enough that a market is emerging to exploit it.
Take Dove's April press release, which read, “One of the biggest threats to the representation of real beauty is artificial intelligence.” The personal care company was celebrating the 20th anniversary of its Campaign for Real Beauty, a marketing initiative that aimed to present women from all walks of life without any digital manipulation. Dove took the opportunity to promise that it would “never use AI to represent real women.” (The main purpose of such statements is, of course, to promote Dove, and in that respect it was successful: the headlines of praise came one after the other.) Around the same time, you may have seen Discover's decidedly anti-AI ad: “You robots are becoming more human by the day!” Jennifer Coolidge tells a call-center worker. “At Discover, everyone can speak to a human agent,” the worker replies.
These are a subsidiary of Unilever and a major credit card company, respectively, so in other words, they may not be organizations we would normally look to for moral clarity, but their ads respond to real fears. And it's not just a corporate ad campaign. New companies are being formed to cater to users disillusioned with generative AI. Cara, a social media and portfolio app for artists, has explicitly prohibited users from publishing AI-generated artworks in its terms of use since its 2023 launch. Users have been flocking to it in recent weeks since news broke that Meta, which owns Instagram, would automatically populate all public posts with AI training data. The app briefly rose to number five on the iOS social network charts, and in a matter of days saw its user base grow from 40,000 to nearly 1 million.
“We want a platform that excludes images from scraping by default, that ensures that datasets are ethically sourced, and that doesn't host AI media until laws are passed to protect artists' work,” Cara founder Jinnah Chan told me. Users seem to want that, too. “Words can't express how good it feels to be here, and to know that what I'm seeing here is human-made,” artist Carla Ortiz said in a June 2 Cara post. The post has been liked 10,900 times so far. (Ortiz is one of the plaintiffs in a recent class-action lawsuit that alleges AI companies infringed on artists' copyrights.)
Her joy at finding a haven in an AI-fatigued internet is perhaps not surprising. As AI-generated content proliferates online, so too do concerns about the quality, ethics, and safety of the technology. Generative AI services are still prone to “hallucinating” and providing false and unreliable information, and could be used to generate fraud and misinformation. They are also trained on the work of unauthorized creators, most of whom are not paid. As such, companies, brands, and creative workers have steadily moved to explicitly promote their products and services as human-made, a bit like the organic food labels that gained traction a few years ago, but for digital labor. Certified 100% AI-free.
Writers and media are posting disclaimers and “no AI” declarations on their blogs and websites. The group Not by AI offers a downloadable badge that anyone can use (it says 264,000 web pages do so now). A classical radio station in Omaha has issued a “no AI” pledge, and the Perth Comic Art Festival has issued a statement banning AI-generated media at its events. Instagram users are using hashtags like “#noai,” “#notai,” and “#noaiart,” a modern version of the #nofilter trend that indicates images are presented without digital alteration. 404 Media, a technology journalism outlet, describes itself as AI-free, “a media by humans, for humans.” In a digital ecosystem that is overwhelmingly controlled by monopolistic tech companies like Google and Meta, and that is rushing to introduce new AI products whether users want them or not, even these small declarations become a flag to protest, to voice discontent, and to rally around other AI skeptics.
These discontents, reflected in the Hollywood writers strike aimed at limiting the use of AI, class-action lawsuits like the one Ortiz is joining, and increased workplace organizing around AI in industries like gaming and journalism, highlight a widespread and sincere desire to put work in human hands, and for quality human-made art, writing, and services.Atlantic We have a corporate partnership with OpenAI. Atlantic It is operated independently from the business divisions.
But when I started following this new trend a few months ago, the first “AI-free” marketing materials I came across were hosted by a tech startup, whose story seemed especially relevant and prescient to me.
Inqwire's site looks a lot like its peers, with its minimalist design and playful branding. In this case, products like Smart Journal “help you identify and explore meaningful topics from your writing.” But rather than touting how it can optimize the latest AI technology, as most tech companies in 2024 would, Inqwire boasts about its complete rejection of AI technology, placing a module in the middle of the homepage and making it bold for emphasis.100% LLM-FreeInquiry Technology do not use We employ large-scale language models (LLMs) and never offer chatbots or conversational interfaces that behave like humans or mimic human experts.”
“I was heartened to see people say, 'I want to pay for the service if the LLM is free,'” Inqwire founder Jill Nephew told me.I I absolutely will,” said the nephew. No LLM required She decided not to use the label for a few reasons: she didn't want to promote a tool that could potentially take people's jobs, she wasn't convinced the LLM was a reliable business solution, and she learned in her early days working at a startup during the first dot-com boom that ultimately what customers wanted was a practical tool whose output she could understand.
Nephew told me that right after college, he got a job working on “black box algorithms” at a hot startup in the '90s called Red Pepper Software. (The company was acquired by PeopleSoft, which was later acquired by Oracle.) The company sold enterprise software to help companies optimize production and delivery schedules. Customers often didn't know why the software produced the results it did, a problem that persists in AI systems today. Nephew spent years refining the system and learned a key lesson, which reflects a problem facing the AI industry today: “People are initially impressed by all the promises of the super-megabrain, but what they really value is something that can be explained, defended, and understood. If you can't understand it, it's not going to happen.”
In other words, Nephew believes the technology is overhyped and underpowered, and that it would be wise to separate your company from the rest before the trend collapses. Similarly, Portland, Oregon-based call center company AnswerConnect has a tagline of “People, not Bots.” In a report the company commissioned from market research firm OnePoll, 78% of respondents said they “want to talk to a real person when contacting a company.” If all of this is true, it makes sense to eschew AI and embrace human workers.
Behind all this AI-Free The label begs a question, one that will resonate even louder as the limitations of generative AI become painfully obvious and the companies responsible for it become more ethically compromised: What is AI-generated diversity? forIn general, humans are preferred over AI or automated systems in customer service. AI art has been widely criticized online, with teenagers scorning it as “baby boomer art.” AI is Better Product, not necessarily: it simply provides moreand for less money. Are you prepared to sacrifice your humanity for it?
In the 2000s, organic and No GMOs The labels were a response to concerns about sustainability, pesticides and factory farming. organic Food labels were meant to signal quality versus inferior products, but the lesson here is, of course, that branding has its limits. organic The labels are expensive and difficult to verify, making them often meaningless, encouraging companies like Whole Foods to sell branding even when it offers little nutritional benefit.
The richest corporations on the planet are promoting generative AI artifacts as cheaper, easier-to-produce alternatives to human art and services. And a few ad campaigns by Doves and Discover won't stop them. We should absolutely build alternative platforms for people who want to put up their badges, ring their AI-free bells, and escape predatory-trained LLMs. But if we want to preserve a human economy for creative goods and services, we're going to have to fight for it, too.
