Training AI systems with more focused data sets can target them to a specific use.
Academic rigor, journalistic flair
Articles on Large language models
Displaying 21 - 40 of 53 articles
Language model AIs seem smart because of how they string words together, but in reality they can’t do anything without many people guiding them every step of the way.
Snapchat’s AI-powered chatbot malfunctioned this week, raising questions of “sentience” among users. As AI becomes increasingly human-like, society must become AI-literate.
Giving AI any degree of executive control could be dangerous for humans.
ChatGPT can be very useful – if you shift how you view it. The first step is to stop thinking of it as a chatty search engine.
In a preprint study, researchers estimate training the model behind ChatGPT would have required somewhere between 210,000 and 700,000 litres of water.
Mona Awad and Paul Tremblay’s lawsuit claims their books were used without their consent. But copyright protection doesn’t apply to ideas – they’ll need to demonstrate the likelihood of economic loss.
To be useful in high-stakes situations, AI needs to understand cause and effect – and the limits of its knowledge.
Public comment could soon swamp government officials and representatives, thanks to AI, but AI could also help spot compelling stories from constituents.
AI can quietly work in the background as a sophisticated technological tool that delivers better results and performance for people.
We humans like to think that our language is original, but we absorb large amounts of it from others and liberally repeat and remix what we hear – just as language AIs do.
Schools are blocking access to ChatGPT on their computers to try to prevent students from cheating. Two experts on academic cheating offer a very different strategy.
Artificial intelligence looks like a political campaign manager’s dream because it could tune its persuasion efforts to millions of people individually – but it could be a nightmare for democracy.
Of the risks posed by AI, overtaking human intelligence isn’t an immediate concern.
One researcher’s experience from a quarter-century ago shows why bias in AI remains a problem – and why the solution isn’t a simple technical fix.
Do you own a literary work that ChatGPT helped you write? Does OpenAI? The legal questions are thorny, and the answers unclear.
There’s a lot of overlap in AI-related concepts. Understanding how they are different from each other, and how they relate, is important.
Generative AI can seem like magic, which makes it both enticing and frightening. Scholars are helping society come to grips with the potential benefits and harms.
Language model AIs are smooth talkers, but you shouldn’t rely on them to make important decisions. That’s because they have trouble telling the difference between a gain and a loss.
Large language models can’t understand language the way humans do because they can’t perceive and make sense of the world.
Top contributors
-
Professor of Information Science, University of Washington
-
Lecturer in Business Analytics, University of Sydney
-
Associate Professor of Computer Science, Florida International University
-
Emeritus Professor of Psychology, Arizona State University
-
Research Assistant Professor of Industrial & Systems Engineering, University of Southern California
-
Professor of AI, Research Group Leader, UNSW Sydney
-
Associate Vice Chancellor, University of Tennessee
-
Associate Professor of Computer Science, Affiliate Professor of Information Science, University of Colorado Boulder
-
Professor of Biomedical Informatics, University of Colorado Anschutz Medical Campus
-
Postdoctoral Research Fellow in Ethics and Societal Implications of Artificial Intelligence, Georgia Institute of Technology
-
Professor of Philosophy and Director, Applied Ethics Center, UMass Boston
-
Professor of Community Information, University of Michigan
-
Doctoral Student in Cognitive Science, University of California, San Diego
-
Director, Enterprise AI and Data Analytics Hub, RMIT University
-
Professor of Computer Science, University of California, Berkeley
