Author: The Conversation

In graduate school, my experimental archaeology professor told a student to create a door socket – the hole in a door frame that a bolt slides into – in a slab of sandstone by pecking at it with a rounded stone. After a couple of weeks, the student presented his results to the class. “I pecked the sandstone about 10,000 times,” he said, “and then it broke.” This kind of experience is known as individual learning. It works through trial and error, with lots of each. Also known as reinforcement learning, it is how children, chimpanzees, crows and AI often learn to do something on…

Few actors are imitated as often as Michael Caine. Even Michael Caine has imitated Michael Caine. His voice has been used in birthday card greetings and has been the source of jokes in various comedy sketches. It is synonymous with a certain type of Britishness. Last week, artificial intelligence (AI) company ElevenLabs announced that Caine has licensed his voice to the company. It will be available on their ElevenReader app, which allows you to listen to any text in a voice of your choosing, as well as being available on their licensing platform, Iconic Marketplace. To understand why Caine’s voice is so iconic (and…

How did life begin on Earth? While scientists have theories, they don’t yet fully understand the precise chemical steps that led to biology, or when the first primitive life forms appeared. But what if Earth’s life did not originate here, instead arriving on meteorites from Mars? It’s not the most favoured theory for life’s origins, but it remains an intriguing hypothesis. Here, we’ll examine the evidence for and against. Timing is a key factor. Mars formed around 4.6 billion years ago, while Earth is slightly younger at 4.54 billion years old. The surfaces of both planets were initially molten, before…

Artificial intelligence (AI) is at a very Chinese time in its life. Recent moves from Chinese AI labs are throwing the dominance of American “frontier labs” such as Google and OpenAI into question. Last week ByteDance, the company behind TikTok, released an AI video-generating tool called Seedance 2.0 which produces high-quality film-like clips from text prompts, with a casual disregard for copyright concerns. This week Anthropic, the US company behind the chatbot Claude, said three Chinese AI labs created thousands of fake accounts to harvest Claude’s answers in a practice called “distillation” which can be used to improve AI models. These events have led…

When a human says an event is “probable” or “likely,” people generally have a shared, if fuzzy, understanding of what that means. But when an AI chatbot like ChatGPT uses the same word, it’s not assessing the odds the way we do, my colleagues and I found. We recently published a study in the journal NPJ Complexity that suggests that, while large language model AIs excel at conversation, they often fail to align with humans when communicating uncertainty. The research focused on words of estimative probability, which include terms like “maybe,” “probably” and “almost certain.” By comparing how AI models and humans map…

It’s 1 a.m. The argument is over, but you keep running it back in your head anyway. You replay the tone, timing and that one sentence that landed wrong. So you open an artificial intelligence (AI) chatbot and type, “Am I right or am I overreacting? What do I say to what they said? What did they mean by XYZ?” Research on attachment, emotion regulation and online discourse helps explain why turning to AI is becoming increasingly popular. The reassurance that it provides, however, can consolidate a one-sided interpretation far too quickly and, ultimately, train expectations that real relationships struggle to meet.…

As a professor of the future of work, the question I get asked most often is whether AI is going to take everyone’s jobs. I hear it from students who worry that their degrees will be obsolete before they graduate. I hear it from office workers watching new tools appear in their software. And I hear it from people working in retail and logistics and hospitality and admin, who all suspect that their jobs put them most at risk. The issue has become a widespread worry in the workplace. And of course, I understand why people are worried. Because for a very…

American artificial intelligence (AI) company Anthropic this month attracted applause – and a surge in users – for clever advertisements poking fun at its competition. In the commercials, an AI assistant awkwardly breaks away mid-conversation to push products such as shoe insoles and dating services. “Ads are coming to AI”, the Super Bowl-tied spots warned, but not to Anthtropic’s own chatbot Claude. The campaign quickly generated buzz because it played to peoples’ worries that inviting advertising into AI platforms, which many of us now rely on – and confide in – risks blurring the line between helpful advice and paid influence. But that anxiety, while understandable, overlooks how advertising already works across much…

“Edge computing”, which was initially developed to make big data processing faster and more secure, has now been combined with AI to offer a cloud-free solution. Everyday connected appliances from dishwashers to cars or smartphones are examples of how this real-time data processing technology operates by letting machine learning models run directly on built-in sensors, cameras, or embedded systems. Homes, offices, farms, hospitals and transportation systems are increasingly embedded with sensors, creating significant opportunities to enhance public safety and quality of life. Indeed, connected devices, also called the Internet of Things (IoT), include temperature and air quality sensors to improve indoor comfort, wearable sensors to monitor patient health, LiDAR and…

Most AI training teaches you how to get outputs. Write a better prompt. Refine your query. Generate content faster. This approach treats AI as a productivity tool and measures success by speed. It misses the point entirely. Critical AI literacy asks different questions. Not “how do I use this?” but “should I use this at all?” Not “how do I make this faster?” but “what am I losing when I do?” AI systems carry biases that most users never see. Researchers analysing the British Newspaper Archive in 2025 found that digitised Victorian newspapers represent less than 20% of what was actually printed. The sample…