Trending Now
We have updated our Privacy Policy and Terms of Use for Eurasia Group and its affiliates, including GZERO Media, to clarify the types of data we collect, how we collect it, how we use data and with whom we share data. By using our website you consent to our Terms and Conditions and Privacy Policy, including the transfer of your personal data to the United States from your country of residence, and our use of cookies described in our Cookie Policy.
{{ subpage.title }}
The world of AI in 2025
1. Models will improve—but how much? There is serious concern that the rate of progress for artificial intelligence is actually slowing down—perhaps there are diminishing returns on large language model investment. Will we achieve AGI—some artificial superintelligence—in 2025? Likely not. If anything, the models could start to underwhelm. If that happens, maybe the hype cycle for AI will begin to twilight. It’ll be up to the industry’s leaders to prove to the public and to investors that the future is bright—even if it requires some patience before the much-hyped awesome future of AI reveals itself.
2. Meet your new assistant: In 2025, you will encounter AI that looks slightly different. Corporate leaders are focusing on multimodality, the idea that chatbots should be able to use text, images, videos, and more all in one interface. But AI executives are already talking about the next great thing: AI agents. These models, in theory, will be able to perform tasks for you — like fill in time cards or order your groceries — not simply chat with you and answer requests. But do we trust AI programs to actually work autonomously? Or should they stay completely responsive to our every line to text?
3. Labor will fight back: Organized labor has already pushed back against the advent of artificial intelligence, especially in Hollywood and the video game industry. Expect that creative workers from all walks of life may get more frustrated with AI companies allegedly ripping off their work and not compensating them for it. That could mean more lawsuits, but also more strikes. Can labor and tech find a happy compromise? It could prove crucial.
4. Trump will refocus regulation: President-elect Donald Trump will take a hands-off approach to regulating artificial intelligence — unless AI companies cross him. Right now, tech leaders from Meta, Amazon, OpenAI, and even Perplexity are donating to Trump’s inaugural fund in order to get on his right side. But we’ll also see the influence of AI executive Elon Musk on the administration: Will he enact his own revenge, or use his position in charge of government efficiency to cut federal contracts for AI?
5. War will get more automated: The government may not fully embrace AI in all of its departments and agencies, but the military almost certainly will. We’ve written extensively about how the Pentagon is embracing AI and firms are lining up to become defense contractors or bolster their existing contracts with additional AI services. And with the AI industry still searching for a profitable business model—ignoring the ethics of it all—why not defense contracting?
2024: The Year of AI
The Trends
1. Beyond chatbots: In 2024, generative AI moved outside of the chatbot interface toward more creative software applications. For GZERO AI, we tested out quite a few — AI avatars, artificial voices, music generators, search engines, and fact-checkers — some were impressive, some not so much. But they showed there’s a real appetite and use for this technology outside of text conversations with ChatGPT and Claude.
2. The chip race: Companies are clamoring to get their hands on as many powerful chips as possible—but they’re expensive: Meta alone revealed it’s spending billions on Nvidia graphics processors. Meanwhile, just about every major US tech company — including Amazon, Apple, Meta, Microsoft, and OpenAI — is trying to reduce their dependence on Nvidia and AMD by designing their own chips. Getting access to cheaper chips more quickly could be the difference between an AI industry that’s profitable or not.
3. Energy crisis: The climate crisis ran head-on into the ambitions of AI companies this year. Google and Microsoft have each admitted that the electricity demands of artificial intelligence are requiring them to reconsider or revise their goals to lower carbon emissions. Companies began talking much more about — and striking deals to acquire — nuclear power, considered a clean energy source due to its lack of carbon emissions.
4. Regulators act: While Congress was deadlocked as usual, unable to pass real AI legislation, state lawmakers made progress. In California, where Gov. Gavin Newsom vetoed a major AI safety bill, he did sign into law a transparency bill for AI-generated content and a bill that requires tech platforms to remove election-related deepfakes. At the federal level, the Biden administration cracked down on chip exports and ordered his agencies to develop AI rules. The US lagged behind Europe on comprehensive AI regulation, but it didn’t totally stand still.
The Moments
5. Europe regulates AI: The European Union’s long-in-the-works Artificial Intelligence Act finally passed into law in the spring, providing the first major law regulating AI and a framework for other governments. This first-mover bill won’t take effect until 2025, but was undoubtedly the year’s major regulatory accomplishment.
6. The United Nations gets the ball rolling: In September, a United Nations advisory body unveiled its own long-awaited plan for AI, outlining specific line-items for the global governance body in order to ensure a “globally inclusive” system. The group called for an office within the Secretariat, a scientific panel on AI, and more.
7. The Nobels: AI research was recognized in a big way this year. The 2024 Nobel Prizes inordinately awarded advances that used or paved the way for artificial intelligence. The Physics prize went to Geoffrey Hinton and John Hopfield for their work on artificial neural networks; and, David Baker, Demis Hassabis, and John M. Jumper won the Chemistry prize for predicting protein structures using AI.
8. AI didn’t ruin the election: While many experts feared that AI would be used to deceive or trick voters in the US election, that largely didn’t come to fruition. The technology was used deceptively — mimicking Joe Biden’s voice on robo-calls, and fake videos of Kamala Harris on social media — but nothing majorly confused voters or tangibly affected the voting process. That doesn’t mean there isn’t possible harm — it just means we dodged a bullet.
The People
9. Sam Altman and Dario Amodei: While the greatest innovations of 2024 weren’t chatbots, they still dominate the industry. And none are bigger than ChatGPT and Claude, led by Sam Altman’s OpenAI and Anthropic’s Dario Amodei. After being fired at the end of 2023, Altman came back to reclaim his company and released new updates to the GPT-4 model as well as the new text-to-video model Sora. Meanwhile, Claude became an industry favorite with its new models, especially Claude 3.5 Sonnet, for its writing and analytical abilities. Both men boosted their companies’ profiles, funding, and position leading the industry in 2024.
10. Jensen Huang and Lisa Su: There is no AI without the chips that power it. That’s what Jensen Huang and Lisa Su, the chief executives of Nvidia and AMD, cemented in 2024. Even with stringent government rules about who they can ship their products to, the two chip design firms have succeeded where legacy chipmakers have failed. And Huang and Su have both guided their once-small companies to glory, becoming irreplaceable mainstays of the industry, leaders that no one can avoid if they want to succeed in AI.
11. Gina Raimondo: The US government has succeeded in keeping the most powerful AI technology out of China’s hands. And that’s thanks to the efforts of US Commerce Secretary Gina Raimondo, who leads the export control initiatives for Joe Biden. Under Raimondo, the government has cut off the sale of powerful computer chips and semiconductor manufacturing equipment to China and its allies. Some chips still find their way in, but if China has anyone to blame for lagging far behind the US, it might be Raimondo.