Tag: language models
-
ChatGPT shows hiring bias against people with disabilities
Language models exhibit hiring bias against people with disabilities, ranking resumes with disability-related achievements lower than those without. However, this bias mirrors existing societal prejudices in human hiring practices. While AI bias can be reduced through simple prompts, addressing human bias is a lot more challenging.
-
Podcast – AI and the Evolution of Medical Thought
In this episode of the AI Grand Rounds podcast, Dr. Adam Rodman shares his unique journey from a historian to a physician deeply interested in the intersection of medicine and artificial intelligence.
-
Building custom GPTs to provide feedback on samples of writing
Introduction “Our results show that i) ChatGPT is capable of generating more detailed feedback that fluently and coherently summarizes students’ performance than human instructors; ii) ChatGPT achieved high agreement with the instructor when assessing the topic of students’ assignments; and iii) ChatGPT could provide feedback on the process of students completing the task, which benefits…
-
AI assistants will be everywhere
https://www.oneusefulthing.org/p/an-ai-haunted-world “In the near future, AIs will work in their own hierarchies of intelligence, all communicating with each other, perhaps mostly autonomously. If you want to grab a bite to eat, it may be that your more “intelligent” premium AI assistant can guess what restaurant you might like based on reasoning about you and your…
-
Link: A Robot the Size of the World
https://www.schneier.com/blog/archives/2023/12/a-robot-the-size-of-the-world.html “This future requires us to see ourselves less as individuals, and more as parts of larger systems. It’s AI as nature, as Gaia—everything as one system. It’s a future more aligned with the Buddhist philosophy of interconnectedness than Western ideas of individuality. (And also with science-fiction dystopias, like Skynet from the Terminator movies.) It…
-
Hallucinations aren’t a problem to be fixed
A few months ago I wrote a post explaining that language models don’t sometimes hallucinate; they always hallucinate. “…every single response is a creative endeavour. It just happens to be the case that most of the responses we get map onto our expectations; we compare the response against our (human) models of reality.” So I…
-
My biased enthusiasm for generative AI, clearly articulated
I understand the serious ethical concerns many have raised about generative AI. These are important issues that deserve thoughtful debate. I believe that I know something about these concerns and I know the critical position I’m meant to take as an academic working at the intersection of education and technology. I’ve spent 15 years in…
-
Podcast: The trouble with AI
https://samharris.org/episode/SE5DD11091B Sam Harris speaks with Stuart Russell and Gary Marcus about recent developments in artificial intelligence and the long-term risks of producing artificial general intelligence (AGI). They discuss the limitations of Deep Learning, the surprising power of narrow AI, ChatGPT, a possible misinformation apocalypse, the problem of instantiating human values, the business model of the…
-
Link: I’m a student. You have no idea how much we’re using ChatGPT.
Via Dave Nicholls’ Substack newsletter. https://www.chronicle.com/article/im-a-student-you-have-no-idea-how-much-were-using-chatgpt “In reality, it’s very easy to use AI to do the lion’s share of the thinking while still submitting work that looks like your own. Once that becomes clear, it follows that massive structural change will be needed if our colleges are going to keep training students to think…
-
Weekly digest (26 Jun – 02 Jul 2023)
I’ve been thinking a lot about practical guidance for academics who are feeling disoriented by ChatGPT and other language models. Some recognise there’s no getting away from this, and are actively experimenting with the technology to try and understand it better. Even if they’re not embracing LLMs with a happy heart, they can see we…
-
The more I use language models, the more I want to use language models.
-
Don’t use assessment design to solve the ChatGPT problem…
Note: These are my first thoughts on this idea, so I may not be articulating this well. Feel free to tell me how wrong I am in the comments 🙂 Don’t use assessment design to solve the ChatGPT problem. Use ChatGPT to solve the assessment design problem. I’m seeing a lot of chatter around the…
-
Assessment idea: Debate ChatGPT
I know a lot of people are concerned about students’ use of ChatGPT, especially when it comes to essays. The question I get asked most often is, “How do I make sure that students don’t cheat?” I always try to reframe the question to something like, “How do I make sure students are learning?” When…
-
Bing allows you to modulate the amount of ‘hallucination’ in your response
Last week I wrote about LLM hallucinations, and how this isn’t the problem that everyone thinks it is. “I expect that soon we’ll see language models with features that allow us to modulate the output in some way. We may want to dial up creativity or serendipity, in which case we’ll see less overlap with…
-
The “problem” of citation in language models
Recent large language models often answer factual questions correctly. But users can’t trust any given claim a model makes without fact-checking, because language models can hallucinate convincing nonsense. In this work we use reinforcement learning from human preferences (RLHP) to train “open-book” QA models that generate answers whilst also citing specific evidence for their claims,…
-
Customising language models with your own data
Large language models like GPT-4 and ChatGPT can generate high-quality text that is useful for many applications, including chatbots, language translation, and content creation. However, these models are limited to the information contained within their training datasets… To solve this problem, we can augment our LLMs with our own custom documents. Dickson, B. (2023, May…
-
ChatGPT for travel plans…
A friend of mine is using ChatGPT to help plan a week-long trip to Paris for him and his family. He says it’s working really well. We need to think of language models as assistants. They’re fallible and will sometimes make poor suggestions. But a lot of the time, they’ll be incredible.
-
Conversation with Dave Nicholls on generative AI
A few weeks ago I had a conversation with Dave Nicholls for his Paradoxa substack, on the topic of large language models and other forms of generative AI. Dave and I wanted to talk about the implications of generative AI, and of ChatGPT in particular, on higher education and professional practice. In addition to our…