The Top AI Stories of 2025: AI Coding, AGI, and More

In 2025, artificial intelligence was less about flashy demonstrations and more about hard questions. What actually works? What breaks unexpectedly? And what are the environmental and economic costs linked to the further development of these systems?

It’s a year in which generative AI has moved from novelty status to mainstream use. Many people have become accustomed to using AI tools at work, getting their answers through AI search, and confiding in chatbots, for better or worse. It was a year in which tech giants promoted their AI agents, and the general public generally seemed uninterested in using them. It also became impossible to ignore AI slop: it was even Merriam-Webster’s word of the year.

Throughout it all, IEEE Spectrum’Coverage of AI has focused on separating signal from noise. Here are the stories that best illustrate the current state of the field.

Alamy

AI coding assistants have gone from novelty to everyday infrastructure, but not all tools are equally capable or reliable. This practical guide to Spectrum Editor-in-Chief Matthew S. Smith evaluates today’s leading AI coding systems, examining where they significantly improve productivity and where they still fall short. The result is a clear view of which tools are worth adopting now and which remain better suited for experimentation.

Amanda Andrade-Rhoades/The Washington Post/Getty Images

While there are concerns about the energy needs of AI, water consumption has become a more discreet but equally urgent problem. This article explains how data centers consume water for cooling, why impacts vary widely by region, and what engineers and policymakers can do to reduce this pressure. Written by Shaolei Ren, AI Sustainability Specialist, and Amy Luers, Head of Sustainability at Microsoft, the article grounds a vocal public debate about data, context, and technical reality.

iStock

When AI systems fail, they don’t fail like humans. This essay, written by legendary cybersecurity guru Bruce Schneier and his frequent collaborator Nathan E. Sanders, explores how machine errors differ from human errors in structure, scale, and predictability. Understanding these differences is key to creating AI systems that can be responsibly deployed in the real world, researchers say.

Christie Hemm Klok

In this insider story, John Dean, co-founder and CEO of WindBorne Systems, tells readers how his team built one of the most technically ambitious AI forecasting systems to date. The company’s approach combines long-duration autonomous weather balloons that ride the wind with a proprietary AI model called WeatherMesh, which sends the balloons high-level instructions about where to go next and analyzes the atmospheric data they collect.

WindBorne’s platform can produce high-resolution forecasts faster, using significantly less computation, and with greater accuracy than conventional physics-based methods. In the article, Dean walks readers through the technical compromises, design decisions, and real-world testing that shaped the system from concept to deployment.

Eddie Guy

This elegantly written article is my favorite of 2025. In it, Spectrum Freelancer Matthew Hutson tackles one of the biggest and most controversial questions in AI today: how to define artificial general intelligence (AGI) and measure progress toward this elusive goal. Drawing on historical context, current debates about benchmarks, and insights from leading researchers, Hutson shows why traditional testing fails and why creating meaningful benchmarks for AGI is so difficult. Along the way, he explores the profound conceptual challenges of comparing artificial and human intelligence.

Bonus: try the test the AIs take to see how smart they are!

IEEE Spectrum

Every year I roll up my sleeves Spectrum’s AI Editor and browse the extensive Stanford AI Index to surface the data that really matters for understanding AI’s advances and pitfalls. The 2025 Visual Summary distills a 400+ page report into a dozen charts that illuminate key trends in the AI ​​economy, energy consumption, geopolitical competition, and public attitudes.