Will AI Take over the World?

by Mallory Knodel
Ahn Young-joon / AP
A screen shows an image for artificial intelligence (AI) at the SK Hynix' booth during a semiconductor exhibition in Seoul, South Korea, Friday, Oct. 16, 2026.

And other questions about the rapidly scaling technology, from environmental impacts to an economic bubble.

OpenAI’s release of ChatGPT in 2022 accelerated the widespread adoption of artificial intelligence (AI) across the globe. Three years later, in 2025, 90 percent of US organizations reported using AI in at least one business function, and the United Nations projected the AI market would hit $4.8 trillion by 2033. But as the industry has grown, so too have people’s concerns.

Data from the 2026 Chicago Council Survey show half of Americans today see AI as a critical threat, and 52 percent say there is not enough regulation of the technology. Recently, people in at least 38 US jurisdictions voted in the midterm elections to limit or ban data centers in their communities. Are there good reasons to be turning on the technology?

Senior Nonresident Fellow for Technology and Governance Mallory Knodel spoke with the Council's Christina Colón about how we got here and what the real risks are for the environment, the US economy, and competition with China.

This interview has been edited for length and clarity.

One thing we’re hearing more about is that AI is dangerous and, in particular, bad for the environment. There are, of course, the AI data centers and then there is also this thought that AI is operating separately from us and the dangers inherent to that. Do you think AI is dangerous? 

It's a really complicated question because it's almost like saying software is dangerous. And absolutely, there is dangerous software out there. There's software that's built to power weapons and software that is for hacking and cybersecurity breaches. But today, a lot of the ways that we're interfacing with AI is really just the way we're interfacing with computers. Even if you're not talking to a chatbot, the software you’re using, the devices you’re using, and the services that power all of that are using AI. It is becoming part of our daily lives. And yes, it can be dangerous—and that can be purposeful or it can be accidental—but it can also be very mundane and very normal.

I think we have to be cautious of taking a very narrow view on something that is now able to do almost anything we ask it to do. In that way, it becomes incumbent on us as people, as the ones in charge, to figure out what's right and what's wrong and to ensure that the AI—the data-driven software—is working for us and in accordance with the social norms and the laws that we've set up in our society. That has always been true with technology. AI is just the latest.

Another concern that people have is the economic impact of this widespread implementation of AI. We hear about a “K-shaped” economy where these tech CEOs are getting very rich from AI while many Americans are struggling. More recently, people have begun to wonder if AI is just a bubble waiting to burst. What do you think? Is AI a bubble? 

The thing about AI as we're dealing with it right now is scale. We're not just talking about a few data points here and there. We are talking about massive volumes of data. Basically, all of human history that can be distilled into data points has been used to build these models. It’s why there aren't that many of them and what makes the people who control those models extraordinarily powerful.

Earlier software was rule-based. You set up the rules, you put in the data, and you got an answer. But these large language models (LLMs) are using piles of data. That data has to sit in these data centers, and that's just not a super efficient way of doing things. It's very expensive to grow and it's going to impact our lives, no question. And so, when folks talk about a bubble, they're talking about some of that material reality. But then they're also talking about the financial situation.

I can certainly tell you that there is not a clear business model for a lot of the AI applications that we see and that we are most impressed by. Sure, there's a subscription, but is that small subscription I'm paying taking into account all of what it really costs? And I know we're not usually including things like externalized costs, but I think when the impact on the environment is so extreme, we must.

There is not a clear business model for a lot of the AI applications that we see and that we are most impressed by. 

When you add all of that up, the bubble appears to be a lot of scale with not a lot of substance and a lot of downstream costs. What worries people is that we're not actually going to see returns on this. And if it's just kind of replacing the software we already use in a lot of ways, then is that really worth all of this investment? Is that really going to be so revolutionary? And is our economy even able to handle that kind of cost for not a lot of gain?

I'm skeptical. I see the way it works, and I don't know that it's as impressive as it ought to be given how much people have invested in it—time, resources, and money.

And this is one area where we think more about global implications. Recently, many Americans voted against building these massive data centers in their communities, which has raised some concern about how that could impact our competitiveness in this “AI race” we’re in with China. What are your thoughts on this? 

Our current moment is entirely defined by what has happened in the past. The harms that we are worried about today are actually pretty longstanding. This is a continuous story. The solutions that are needed today feel so urgent and so galvanizing because the same people had their jobs taken from them by the internet 15 to 20 years ago, right? And we didn't rectify that.

Whole industries were disrupted—hospitality, travel, local transport, a variety of things. Point at an industry and, at some point, some tech company has tried to or has done a good job of disrupting it. It is extractive. And I think this is being expressed through the data center protests because it's something physical coming to their communities. Often this is all very abstract. Tech is in the cloud; it is not in our hands. But the data centers are proximal. It's something that we can demonstrate our opposition to.

I don't think anybody in this movement would agree that the data center just moving to another county or another country is a win, though. The idea of the movement is to slow it down. We're asking leaders not just to compete on a global scale for the economy and for financial security, but to help us through this moment and ensure that there's a plan for jobs and for dealing with the climate implications. And we aren't seeing that kind of leadership.

That demand for leadership also gets at the safety and security aspect of AI and this question of whether there can be a global governance system. In fact, one of the most frequently asked questions is what happens when AI goes rogue. 

It's important to remember that this is the same thing we had before but scaled up. We've had massive data breaches and real global events with the internet, and our cybersecurity mechanisms are relatively intact. But cybersecurity is one of those things that is as multifaceted and complicated as the threats.

There are different ways to set norms at a global governance level. And there are aspects of the internet that are also just their own sort of universe or governance structure. So, on the one hand, I think it's really important to follow these incidents very closely. But I think it's important to respond by increasing the capacity of existing institutions that are there to harden our systems at every layer and in all aspects, because that is what actual cybersecurity takes. Today, it's AI but it could be something else in the future. So, it’s more about building the institutions and the infrastructure than about individual responses.

For example, there has long been reticence to legislate around privacy. All of those massive piles of data that fed these LLMs and made them so powerful wouldn't have been as easy to make if we had proper privacy protections. We allowed companies to take our personal data and make their own products out of it, and now they’re selling it back to us. And there has been so much excitement around AI and its use and what it can do that it has sort of detracted from the goal of getting more privacy protections legislated.

So, yes, AI going rogue is a real threat. but I think we have mechanisms to deal with it. They just need to be scaled appropriately to the threat.

So, will AI take over the world?

I don't know the future, but I think I'm close enough to some of these things that I don't think it is productive to worry about that right now. The real issues that we are facing together are impacting people every day. We need to focus on those because again, where we are today is a product of the past, and the negligence or intentional decision to not get involved and to not regulate tech has led us to this moment of precarity and fear. We shouldn't be worried so far in the future that we forget that there are real ways that we can change this technology today and that nothing is inevitable.

What we're charged with today is understanding how the technology actually works and what its limitations are rather than just its promises. And then we need to try to meet people where they're at when they're asking for governance and leadership, both at the global and national level, on AI in our society.


The Chicago Council on Global Affairs is an independent, nonpartisan organization and does not take institutional positions. The views and opinions expressed in this commentary are solely those of the author.

About the Expert
Senior Nonresident Fellow, Technology and Governance, Chicago Council on Global Affairs
Mallory Knodel headshot
Mallory Knodel is an expert on how technology shapes power, security, and human rights across the globe. She chairs the Internet Research Task Force’s Human Rights Protocol Considerations research group and previously served as chief technologist at the Center for Democracy & Technology in Washington, D.C.
Mallory Knodel headshot
About the Interviewer
Director, Editorial and Digital Content
Headshot for Christina Colón.
Christina Colón joined the Chicago Council on Global Affairs in 2023. Prior to joining the Council, Colón was the associate editor of Sojourners magazine. She has also served as the communications manager of the Nonprofit Association of Washington and as an editor at Global Press Journal.
Headshot for Christina Colón.

Related Content

President Donald Trump, left, speaks with Meta CEO Mark Zuckerberg as Dario Amodei, CEO and co-founder of Anthropic as he speaks with reporters outside the West Wing of the White House, Tuesday, Sept. 29, 2026, in Washington. Technology
Alex Brandon / AP
Why Big Tech is Betting on the US Midterms
Analysis Mallory Knodel
With federal regulation hanging in the balance, powerful AI companies are backing candidates on both sides of the aisle in a bid for control.