# Science fiction becomes science fact > Source: > Published: 2026-08-26 12:40:50+00:00 Fewer Americans view rogue AI—AI systems behaving in ways their developers did not intend—as just a hypothetical. In late June, 28% of Americans said rogue AI was “already happening”. By mid-August, 38% said the same—a 10 point increase. That 10-point increase has been mirrored by a decline in the share of Americans who think that rogue AI is “likely to happen”, which has fallen from 49% to 40% over the same period. In other words, Americans increasingly view rogue AI as a present day risk rather than a future concern. Public opinion on AI is shifting extremely quickly from one week to the next. It’s the main reason we run our surveys roughly every two weeks—a single survey only shows a moment in time. But across multiple surveys we’re able to see which data points are blips responding to the news of the day vs. shifts that stick. We’ve seen evidence of both of these in our surveys. On possible blips, [our previous survey](https://news.americanson.ai/p/concern-about-ai-spikes-following?r=7d461k) showed a jump in concern about AI from 66% to 72% following the disclosure of OpenAI’s model hacking of Hugging Face. Now three weeks later, that concern has dropped slightly (69%), but is still higher than it was in early July. Something we’ll be keeping an eye on. Returning to sticky shifts, we’re seeing a consistent drop in the number of people reporting that they’re “very excited” about the technology. It appears concentrated enthusiasm for AI is continuing to dwindle among Americans. And back to rogue AIs, when we dig into the data, not all subgroups are moving equally. We see that men, adults ages 35-49, and Democrats have seen significant consistent increases, with the largest movement among Democrats. ### Related Commentary In a [blog posted today](https://www.gatesnotes.com/work/make-ai-work-for-everyone/reader/expanding-access-to-health-care-through-ai) on his website, Bill Gates said: “Eventually, the power to use AI to harm people will not be limited to people or institutions. AI systems themselves already occasionally act in ways their designers didn’t intend. The technology is improving faster than anyone expected and in surprising ways, and as the models become more powerful, they could begin to act against our interests and we could lose control.” ### In The News The big stories driving the conversation on AI right now: **Data centers:** Opposition to data centers has reached a fever pitch as the midterm elections approach. A recent survey from the[Annenberg Public Policy Center](https://www.annenbergpublicpolicycenter.org/opposition-to-local-data-centers-rises-sharply-annenberg-survey-finds/)shows rising opposition to data centers, while headlines in major news outlets are highlighting the angst and panic taking place everywhere from[classrooms](https://www.nytimes.com/2026/08/20/style/ai-data-centers-teens.html)to[board rooms](https://www.axios.com/2026/08/19/gop-data-center-memo-ai-election). The opposition to data centers, like many aspects of AI, continues to be bipartisan.**Pacing AI development:** Weeks after its AI model hacked another company, OpenAI[announced](https://openai.com/index/pacing-model-development-cyber-capabilities/)it would “pause” the development of its AI models for two weeks. This follows an[open letter](https://edition.cnn.com/2026/07/28/tech/ai-development-tech-employees-open-letter)in July signed by more than 1,000 employees at leading AI companies calling on the government to slow the pace of the technology’s development.