ETH Zurich, EPFL, and Stanford HAI formalize human-centered AI partnership
Featured

ETH Zurich, EPFL, and Stanford HAI formalize human-centered AI partnership

ETH Zurich, EPFL, and Stanford's Institute for Human-Centered Artificial Intelligence have announced a strategic collaboration to advance AI research and education, focusing on open, large-scale foundation models and their societal impact. This partnership aims to strengthen academic leadership in AI, emphasizing human-centered approaches and responsible deployment, while countering the influence of commercially driven AI development.

What do researchers acknowledge ChatGPT for in their papers?

What do researchers acknowledge ChatGPT for in their papers?

The use of large language models (LLMs) like ChatGPT is increasingly common in academic research, primarily for tasks such as text editing and proofreading, with 80% of researchers acknowledging its use for these purposes. Despite the potential for more complex applications, such as coding or data analysis, these are less frequently reported, and many researchers may not disclose LLM use due to unclear journal policies or the perception that certain tasks do not require acknowledgment.

Google's year in review: 8 areas with research breakthroughs in 2025

Google's year in review: 8 areas with research breakthroughs in 2025

In 2025, Google achieved significant advancements in AI research with models like Gemini 3 and Gemma 3, enhancing AI's reasoning, multimodality, and efficiency, which led to new products and features across its portfolio. These developments have not only transformed Google's products but also contributed to scientific progress and global challenges, with a strong emphasis on responsible AI development and collaboration.

Bias in AI: Examples and 6 Ways to Fix it in 2026

Bias in AI: Examples and 6 Ways to Fix it in 2026

Interest in AI is growing among businesses due to its benefits, but concerns about biases in AI technology persist. Studies show that AI models often replicate societal biases, such as racial and gender stereotypes, due to imbalanced training data, leading to significant ethical and social impacts, including discrimination in hiring and wrongful arrests.

Empowering educators with AI innovation and insights

Empowering educators with AI innovation and insights

Microsoft has announced the general availability of Copilot Chat for students aged 13 and older, alongside new AI features for educators in Microsoft 365 Copilot and Copilot+ PCs. The 2025 AI in Education Report highlights significant growth in AI usage among educators, though many still lack confidence in using AI effectively, emphasizing the need for training and resources to prepare students and educators for an AI-powered future.

Making AI work for schools

Making AI work for schools

The rapid integration of AI in education presents both opportunities and challenges, with the global AI-in-education market projected to grow significantly. To ensure responsible adoption, education leaders and stakeholders must focus on transparency, accountability, and continuous monitoring, while implementing structured pilot programs and audits to safeguard equity and trust in AI deployments.

Students are using AI apps to summarise whole books in minutes. Experts are alarmed

Students are using AI apps to summarise whole books in minutes. Experts are alarmed

AI-powered summary tools are becoming popular among students for their ability to condense lengthy texts into brief overviews, but they raise concerns about declining literacy rates and the potential loss of critical reading skills. While these tools offer convenience, they often produce unreliable summaries that may not align with educational expectations, and their increasing use could financially harm authors by reducing book sales.