Author: The Conversation

OpenAI’s latest research paper diagnoses exactly why ChatGPT and other large language models can make things up – known in the world of artificial intelligence as “hallucination”. It also reveals why the problem may be unfixable, at least as far as consumers are concerned. The paper provides the most rigorous mathematical explanation yet for why these models confidently state falsehoods. It demonstrates that these aren’t just an unfortunate side effect of the way that AIs are currently trained, but are mathematically inevitable. The issue can partly be explained by mistakes in the underlying data used to train the AIs. But using mathematical analysis…

Imagine adjusting the temperature of the air conditioning or skipping a song in your car, not by fiddling with a screen or voice command, but simply by swiping your hand across the fabric of your seatbelt. It sounds futuristic, but this is the direction automotive design could be shifting towards — away from screens and buttons, and towards multi-touch textiles that sense your gestures and respond to them. I am an interaction design professor and director of a research lab that develops smart textile technology. These textiles can transform how people interact with everyday objects and materials, including car interiors,…

Social companion robots are no longer just science fiction. In classrooms, libraries and homes, these small machines are designed to read stories, play games or offer comfort to children. They promise to support learning and companionship, yet their role in family life often extends beyond their original purpose. In our recent study of families in Canada and the United States, we found that even after a children’s reading robot “retired” or was no longer in active and regular use, most households chose to keep it — treating it less like a gadget and more like a member of the family. Luka…

Artificial intelligence (AI) is reshaping how students write essays, practise languages and complete assignments. Teachers are also experimenting with AI for lesson planning, grading and feedback. The pace is so fast that schools, universities and policymakers are struggling to keep up. What often gets overlooked in this rush is a basic question: how are students and teachers actually learning to use AI? Right now, most of this learning happens informally. Students trade advice on TikTok or Discord, or even ask ChatGPT for instructions. Teachers swap tips in staff rooms or glean information from LinkedIn discussions. These networks spread knowledge quickly but…

You’re introduced to someone and your attention catches on their eyes. They might be a rich, earthy brown, a pale blue, or the rare green that shifts with every flicker of light. Eyes have a way of holding us, of sparking recognition or curiosity before a single word is spoken. They are often the first thing we notice about someone, and sometimes the feature we remember most. Across the world, human eyes span a wide palette. Brown is by far the most common shade, especially in Africa and Asia, while blue is most often seen in northern and eastern Europe.…

Attempts at communicating what generative artificial intelligence (AI) is and what it does have produced a range of metaphors and analogies. From a “black box” to “autocomplete on steroids”, a “parrot”, and even a pair of “sneakers”, the goal is to make the understanding of a complex piece of technology accessible by grounding it in everyday experiences – even if the resulting comparison is often oversimplified or misleading. One increasingly widespread analogy describes generative AI as a “calculator for words”. Popularised in part by the chief executive of OpenAI, Sam Altman, the calculator comparison suggests that much like the familiar plastic…

Back in the 2000s, the American pharmaceutical firm Wyeth was sued by thousands of women who had developed breast cancer after taking its hormone replacement drugs. Court filings revealed the role of “dozens of ghostwritten reviews and commentaries published in medical journals and supplements being used to promote unproven benefits and downplay harms” related to the drugs. Wyeth, which was taken over by Pfizer in 2009, had paid a medical communications firm to produce these articles, which were published under the bylines of leading doctors in the field (with their consent). Any medical professionals reading these articles and relying on them for prescription advice…

As another school year returns, large language models (LLMs) present difficult questions around learning, thinking, plagiarism and authorship for educators. New approaches to assignments and assessment are required. Student papers that use LLM technology require additional labour on many fronts. Professors have expressed frustration, worry and anxiety. As an assistant professor of English whose research has focused on the histories of writing and how it’s taught, I have been involved in many discussions at institutions of higher learning about this topic. The immediate issue of LLMs in the classroom points to a larger reality. For too long, instructors and universities have been treating students’…

When you chat with ChatGPT, it often feels like you’re talking to someone polite, engaged and responsive. It nods in all the right places, mirrors your wording and seems eager to keep the exchange flowing. But is this really what human conversation sounds like? Our new study shows that while ChatGPT plausibly imitates dialogue, it does so in a way that is stereotypical rather than unique. Every conversation has quirks. When two family members talk on the phone, they don’t just exchange information — they reuse each other’s words, rework them creatively, interrupt, disagree, joke, banter or wander off-topic. They do so because human talk is naturally…

Artificial intelligence (AI) is growing fast, and so is the number of computers that power it. Behind the scenes, this rapid growth is putting a huge strain on the data centres that run AI models. These facilities are using more energy than ever. AI models are getting larger and more complex. Today’s most advanced systems have billions of parameters, the numerical values derived from training data, and run across thousands of computer chips. To keep up, companies have responded by adding more hardware, more chips, more memory and more powerful networks. This brute force approach has helped AI make big…