Data science
fromTheregister
2 hours agoUK National Data Library plan needs work, study finds
The UK's National Data Library needs improved dataset accessibility to support AI development and meaningful analysis.
Meta is working on two proprietary frontier models: Avocado, a large language model, and Mango, a multimedia file generator. The open-source variants are expected to be made available at a later date.
Modern scientific societies are increasingly vulnerable due to their dependence on membership fees and journal subscriptions, which are being challenged by the rise of virtual networking and open-access publishing.
Librarians have been actively collaborating and talking about it almost every day, whether it's creating tutorials and digital learning objectives or thinking about the conversations to have with instructors. It can feel like cognitive dissonance to be actively working with AI on a regular basis and also saying we're constantly thinking about the harms and the biases.
They're trying to get as many eyes on [the Epstein Files] and as much public awareness, knowledge, and understanding of this as possible. They built something that the public can use directly, rather than having it be intermediated by journalists, basically having it be in a format that so many people use in their everyday life.
While AI tools are lowering the barrier to development, the gap between speed and manageability is growing. In just over a year and a half, AI code assistants have grown from an experiment to an integral part of modern development environments. They are driving strong productivity growth, but organizations are not keeping up with the associated security and governance issues.
Bias risks: AI can amplify inequalities, like mislabeling non-native English writing as AI-generated. Privacy concerns: Schools face rising cyberattacks, and data misuse risks are high. Accountability: Human oversight is crucial to prevent over-reliance on AI.
Within a couple of years of ChatGPT coming out, I had come to rely on the artificial-intelligence tool, for my work as a professor of plant sciences at the University of Cologne in Germany. Having signed up for OpenAI's subscription plan, ChatGPT Plus, I used it as an assistant every day - to write e-mails, draft course descriptions, structure grant applications, revise publications, prepare lectures, create exams and analyse student responses, and even as an interactive tool as part of my teaching.
His message is clear: our world is built on abundant energy, around 80% of which has come from fossil fuels over the past 50 years. Because supplies are limited, energy consumption will peak in decades - sooner if humans attempt to limit climate change. To keep global warming below 1.5 °C by 2100, the use of fossil fuels must fall by 5-8% each year - a pace that is too fast for low-carbon energy to keep up with.
I'm less interested in topics than in questions, and I'm less interested in publishing than I am in curation. When I've testified before Congress or dealt with an appropriations bill or a budget negotiation, this question, of what is the return on investments when you're doing R&D, comes up quite often. It's been asked by economists in very formal ways since at least the 1950s, but the data and the methods that were available were really not very strong.
A few years ago, I put together what I felt was a truly innovative concept, which I presented in a conference poster at an international meeting in my field. After the presentation, I spoke to another early-career scientist about my work and how it might apply to their findings. Two years later, they scooped me by publishing a preprint paper that presented my idea, with many of the same verbal formulations and an identical flow of ideas, without any acknowledgement or attribution to my work.
The org revealed the new partnerships in a post celebrating its 25th birthday, and which points out it is among the world's ten most-visited websites, and the only one to be run by a nonprofit. The post notes that 250,000 editors work on at least one Wikipedia article each month, and that editors make 324 changes each minute as they contribute to the 65 million-plus articles the site contains. 1.5 billion unique devices reach Wikipedia each month.
In 2023, Australia abandoned its expensive and bureaucratic scholar-led research-assessment programme. New Zealand followed suit soon after. The hope, according to a transition plan unveiled by the Australian federal government's Department of Education and the research sector, was to find a "more modern, data-driven approach". In the United Kingdom, where financial pressures on universities are especially acute, there are similar calls to reform the Research Excellence Framework (REF), the country's performance-based research-funding system.
Drawing on more than 22,000 LLM prompts designed to reflect the kind of questions people would ask artificial intelligence (AI)-powered chatbots, such as, "How do I apply for universal credit?", the data raises concerns about whether chatbots can be trusted to give accurate information about government services. The publication of the research follows the UK government's announcement of partnerships with Meta and Anthropic at the end of January 2026 to develop AI-powered assistants for navigating public services.