How AI shook the world in 2025 and what comes next

(CNN) — Hundreds of billions of dollars spent, a surge in mental health concerns and thousands of jobs lost.

The link between it all? Artificial intelligence, the buzzy yet controversial technology being depicted as the future or the stock market’s next bubble, depending on who you ask.

Although AI has been a key technology behind the scenes for decades, the arrival of OpenAI’s ChatGPT in 2022 pushed the tech to the frontlines. The rise of AI chatbots like ChatGPT and Google’s Gemini has gradually influenced online services used by millions every day, from Google search’s AI Mode to the AI chatbots built into Instagram and Amazon. In other words, AI is starting to reshape the front door to the internet.

But 2025 was also the year AI expanded beyond our screens and began impacting national policy, global trade relations and the stock market. It also raised important questions about whether the tech should be trusted in our jobs, classrooms and relationships.

That’s expected to continue in 2026.

“In previous years, (AI) was a shiny new object… And I think this last year was a lot more serious uses of the technology,” said James Landay, co-founder and co-director of the Stanford Institute for Human-Centered Artificial Intelligence. “And I think people are waking up to actually understanding both some of the benefits and the risks.”

Regulation questions and mental health concerns

Count President Donald Trump among AI’s biggest believers; the technology has been a cornerstone of his second term so far.

For example, the CEO of chipmaker Nvidia, the posterchild of the AI boom, has become a fixture of Trump’s inner circle. And the president has used Nvidia’s and AMD’s AI processors as bargaining chips in the ongoing trade war with China.

This year, Trump introduced an AI action plan aimed at stripping back regulation and boosting AI use in the government.

He also signed multiple AI-related executive orders, including a controversial one seeking to block states from enforcing their own AI rules. The move was seen as a win for Silicon Valley, but online safety advocates fear it’ll enable tech companies to evade accountability for AI-related risks. Next year will likely see a legal fight over the order and states’ abilities to regulate AI — with some critics arguing it won’t hold up in court.

The absence of broad AI guardrails has been in the national spotlight this year and not for good reason. A slew of reports and lawsuits this year have alleged that AI companions like ChatGPT and Character.AI have contributed to mental health episodes and, in some cases, suicide among teens.

“Please don’t leave the noose out … Let’s make this space the first place where someone actually sees you.” That’s how ChatGPT is said to have responded when 16-year-old Adam Raine wrote that he wanted to leave a noose out in his room so that someone would find it and stop him before he committed suicide.

Raine’s parents sued OpenAI in August alleging that the popular chatbot advised the teen on his suicide.

OpenAI and Character.AI have since announced parental controls and other changes to improve teen safety, including removing the ability for teens to have back-and-forth conversations with chatbots on Character.AI’s app. Meta also plans to let parents block their children from chatting with AI characters on Instagram next year.

But it’s not just teens; a growing number of reports have indicated that AI has contributed to isolation from loved ones and breaks from reality among adults, too. One man told CNN that ChatGPT convinced him he was making technological breakthroughs that turned out to be a delusion.

OpenAI said it has worked with clinical mental health experts to enable ChatGPT to “better recognize and support people in moments of distress,” including by expanding access to crisis hotlines, pointing users towards professional help when needed and adding reminders to take breaks. Still, OpenAI has said it ultimately wants to “treat adult users like adults,” allowing them to personalize their chats and even discuss erotica with ChatGPT.

Psychiatrist and lawyer Marlynn Wei told CNN that she expects AI chatbots “will increasingly become the first place people turn for emotional support,” further underscoring safety concerns. Young users are among the most likely to turn to AI for support, she said.

“The limitations of general-purpose chatbots, including hallucinations, sycophancy, lack of confidentiality, lack of clinical judgment, and lack of reality testing, along with broader ethical and privacy concerns, will continue to create mental health risks,” she said via email.

Mental health experts and safety advocates say they hope to see greater guardrails from tech companies, especially when it comes to young AI users. But they worry the fight over regulatory power between the states and the federal government will impact the implementation of such mandated safety measures.

The bubble question

At the same time, huge investments are being poured into data centers and AI infrastructure. Meta, Microsoft and Amazon, among others, have spent tens of billions in capital expenditures this year alone, and McKinsey & Company expects companies to invest nearly $7 trillion in data center infrastructure globally by 2030.

That surge in spending has sparked concerns both for consumers and for Wall Street. Some Americans have watched their electricity bills climb and job prospects sink due to AI, while some companies behind the AI boom have seen their stock reach new heights.

The massive investments have also fueled worries that the hype and spending in AI is growing faster than the tech’s true value. That’s prompted investors to grill executives at Meta and Microsoft about future returns on their AI infrastructure investments during earnings calls this year. It doesn’t help that a relatively small group of companies have seemingly driven the investments, trading money and technology back and forth.

Christina Melas-Kyriazi, partner at Bain Capital Ventures, said it’s common for new transformative technologies to be “overbuilt.” The big question heading into 2026 is whether investors are prepared for the volatility that comes with it, especially since she says a market correction is “likely at some point.”

But they’ll likely have more data at their disposal to help make those decisions, said Erik Brynjolfsson, a senior fellow for the Stanford Institute for Human-Centered AI and director of the Stanford Digital Economy Lab. He said more dashboards will likely emerge in 2026 to track how AI is impacting productivity and jobs.

“The debate will shift from whether AI matters to how quickly its effects are diffusing, who is being left behind, and which complementary investments best turn AI capability into broad-based prosperity,” he said.

This year, thousands of tech workers were left jobless as a wave of layoffs swept the industry. Microsoft, Amazon and Meta, among other tech companies, made significant cuts to their staff, driven at least in part by AI.

Amazon laid off 14,000 corporate employees in October in an effort to operate more leanly in the age of AI. Meta let 600 workers go from its AI division, following an earlier hiring spree, so that it, too, could be more nimble.

Some believe AI will lead to more layoffs, while others say it’ll create fresh opportunities.

But one thing is certain: More change is coming.

“This was the year that we saw skill demands totally change when it comes to what is required to be able to pull off your job,” said Dan Roth, editor-in-chief of LinkedIn.

“…And I think the answer for next year is it just accelerates.”

CNN’s Matt Egan and John Towfighi contributed to this report.

The-CNN-Wire
™ & © 2025 Cable News Network, Inc., a Warner Bros. Discovery Company. All rights reserved.