US: 16% see AI benefit, 49% use chatbots
Photo by Etienne Girardet on Unsplash
We’re officially in the “AI is everywhere, and nobody trusts it” phase of the hype cycle.
Why People Hate AI (Or Just Distrust It)
That 16% number? It’s depressingly low, but hardly surprising. People see chatbots, deepfakes, and job displacement, not robot butlers or miracle cures. The industry has done a terrible job selling the dream, mostly because the dream still feels decades away for the average person.
I’ve seen this play out with every new tech. VR was going to change everything. NFTs were the future of ownership. Remember when the metaverse was the next internet? Each time, the public shrugged, and rightly so. They’re waiting for something tangible, something genuinely useful that isn’t also vaguely terrifying.
The climb to 49% chatbot use is an interesting counterpoint. It shows people are using AI, even if they don’t like it. They’re typing prompts into ChatGPT for homework, asking Copilot to summarize emails, or getting customer service via some bot with a suspiciously human name. Convenience wins, even over skepticism. That’s a classic human trait: complain loudly, then use the thing anyway because it saves five minutes.
The gap between use and trust is a chasm. It’s the same gap that exists for social media, for big tech companies, for most institutions, really. We’ve been burned too many times by promises of connection and innovation that ended up delivering data breaches and algorithmic manipulation. Why would AI be any different?
The PR Problem That Won’t Die
Tech companies are terrible at public relations when it comes to genuinely new, disruptive tech. They always lead with the most futuristic, least relatable applications, or worse, they sound like a Silicon Valley cult. Remember when Google Glass was going to be ubiquitous? It became a meme and then a footnote. The public doesn’t care about “democratizing access to information” when the information is often wrong, biased, or hallucinated. They care about their job, their privacy, and not being replaced by a sophisticated autocomplete.
The narrative around AI has been hijacked by two extremes: the utopian dreamers and the apocalyptic fearmongers. There’s very little middle ground, very little sober assessment of practical, immediate benefits that don’t also carry massive ethical baggage. When Sam Altman goes on tours talking about the potential for superintelligence, while simultaneously asking for government regulation, it creates cognitive dissonance. It’s like a car manufacturer touting a self-driving car while also warning it might spontaneously combust.
Reddit, naturally, has a field day with this. You see comments like “Of course, nobody trusts it, it’s just a fancy plagiarism machine,” or “AI is going to take all our jobs and then tell us why it’s a good thing.” The skepticism is baked in, honed by years of watching tech companies “innovate” their way into new ways of extracting value from users. The top posts often involve stories of AI failing spectacularly, or doing something creepy, or just being incredibly stupid. It feeds into the existing distrust.
Part of the problem is the term “AI” itself. It’s too broad, too nebulous. It covers everything from the algorithms that recommend your next Netflix binge to the models predicting protein folding. To the average person, “AI” means ChatGPT, or the uncanny valley images it generates, or the news stories about it writing movie scripts. It doesn’t mean the complex mathematical models running in the background of their lives.
AI’s Perceived Threats vs. Real World Use
The concerns aren’t baseless. Job displacement is a very real threat, even if the industry tries to spin it as “job augmentation.” We’ve already seen how automation impacts blue-collar jobs; now white-collar roles are in the crosshairs. Creative fields, coding, journalism – nowhere feels safe. And frankly, telling a journalist their job will be “augmented” by an AI that can write a passable article in seconds feels a lot like telling a horse it’s going to be “augmented” by a car.
Then there’s the disinformation problem, amplified by generative AI. Deepfakes are just the beginning. Imagine a world where every piece of digital evidence can be fabricated, every voice cloned, every video manipulated. That’s not some far-off sci-fi dystopia; it’s happening now. The tools are becoming ridiculously accessible. This erodes trust in everything, not just AI. It makes people question what’s real, and that’s a dangerous path for any society.
The privacy implications are also massive. Every interaction with a chatbot, every prompt, every piece of data fed into these models contributes to a vast training set. Who owns that data? How is it used? How is it secured? These are questions the industry often skirts, or answers with vague assurances. We’ve heard those assurances before, usually right before a massive data breach.
Public Perception Trends: A Comparative Look
Let’s look at how public perception of AI compares to other disruptive technologies at similar points in their adoption curve. It’s not a direct comparison, but the patterns are telling.
| Technology | Early Adoption (%) | Public Trust/Benefit Perception (Early Stage) | Key Concerns (Early Stage) | Current State of Trust |
|---|---|---|---|---|
| Internet (mid-90s) | ~10-20% | High enthusiasm, revolutionary | “Information overload,” pornography, scams | High, but with significant privacy/disinfo concerns |
| Social Media (mid-00s) | ~20-30% | High enthusiasm, connecting people | Privacy, “waste of time” | Low, significant ethical and mental health concerns |
| Smartphones (late-00s) | ~30-40% | High enthusiasm, convenience | Addiction, data usage | High, but with privacy and screen time concerns |
| AI (current) | ~49% (chatbot use) | Very low (16% benefit society) | Job loss, misinformation, ethics, existential | Very low, deeply skeptical |
The current perception of AI is starkly different from previous tech waves. The internet was seen as an unmitigated good. Social media and smartphones, while having early critics, were largely embraced as net positives. AI, however, is being met with a level of suspicion that feels almost unprecedented for a technology with such widespread deployment. People are using it, yes, but they’re doing so with one eye open and their hand on their wallet.
The Regulatory Labyrinth and Corporate Responsibility
Governments are scrambling to catch up, but they’re always behind. The EU is taking a more proactive approach with the AI Act, attempting to classify and regulate AI based on risk. The US is still largely in a “let’s study it” phase, with various executive orders and voluntary commitments from tech giants. It’s a classic example of technology moving at warp speed while policy crawls.
The problem with regulation is two-fold: it can stifle innovation if too heavy-handed, or it can be completely ineffective if too vague or reactive. And who exactly is supposed to enforce these regulations? Do we trust government agencies, many of whom still rely on fax machines, to understand and police cutting-edge AI models? It’s a bit like asking a horse and buggy manufacturer to regulate self-driving cars.
Corporate responsibility, or the lack thereof, fuels much of the public’s distrust. Companies rush to market with products, often prioritizing speed over safety or ethical considerations. The “move fast and break things” mantra, once celebrated, now sounds terrifying when applied to something as powerful and potentially destructive as advanced AI. We’ve seen the consequences with social media: filter bubbles, echo chambers, radicalization. AI could supercharge all of that.
I’ve covered enough product launches to know the drill. A CEO gets on stage, talks about how their new widget will “change the world,” and then sidesteps every single question about privacy, data security, or potential misuse. They’re selling a vision, not a product with known flaws and significant risks. The public sees through that now. They’re not buying the hype quite as readily.
Who Benefits from AI, Really?
This is a crucial question. If only 16% of Americans think AI will benefit society, then who do they think it will benefit? The common Reddit answer: “The rich,” “corporations,” or “the people who own the AI.” It’s an increasingly class-conscious perspective, where tech advancements are viewed through the lens of who profits and who is marginalized.
If AI primarily leads to massive gains in productivity for a few large companies, allowing them to cut labor costs and consolidate power, then society as a whole won’t see much benefit. They’ll see job losses and increased corporate wealth. The argument for AI benefiting everyone often hinges on trickle-down economics, which has a spotty record at best.
We need to see tangible benefits for the average person, not just abstract promises. Tools that genuinely make life easier, safer, or more fulfilling without extracting an exorbitant price in data or privacy. We need to see AI used for things like medical diagnostics, climate modeling, or genuine scientific breakthroughs, not just better ad targeting or more sophisticated spam.
Having covered this industry for a decade, I’ve seen the pendulum swing between techno-optimism and dystopian dread multiple times. Each time, the truth lands somewhere in the middle, but often with more negative consequences for the general public than the initial hype suggests. This time feels different, though. The scale of AI’s potential impact is so vast that the stakes feel higher.
The Path to Rebuilding Trust (If There Is One)
Rebuilding trust will be an uphill battle, potentially an impossible one. It requires transparency, accountability, and a genuine shift in corporate priorities. It means moving beyond the “move fast and break things” mentality and embracing a “move cautiously and build trust” approach. Good luck selling that to venture capitalists.
One step is clear communication about limitations and risks. Stop calling these systems “intelligent” when they’re essentially sophisticated pattern-matchers. Acknowledge that they hallucinate, that they can be biased, that they are tools, not sentient beings. Honesty goes a long way.
Another step is meaningful regulation that protects individuals without stifling genuine innovation. This means global cooperation, because AI doesn’t respect national borders. It means holding companies accountable when their AI systems cause harm. It means prioritizing ethical development from the outset, not as an afterthought.
Finally, we need to see benefits accrue to a wider segment of society. This might involve exploring concepts like universal basic income to offset job displacement, or investing in education and retraining programs on a massive scale. If AI is going to create unprecedented wealth, there needs to be a plan for how that wealth is distributed, or at least how those left behind are supported. Otherwise, that 16% number will only shrink further. It’s a long shot, though. The tech industry isn’t exactly known for its altruism.
