Perjantai 🇫🇮 Föstudagur 🇮🇸

I had an intriguing chat with AI today. After teaching a class to my beginner student, I asked chatGPT to create a mnemonic story for me using the words I wanted to memorise. These were rescind, rebuke, respite, rebuff, and rebut. I was astonished, baffled, and simultaneously frightened by the rate at which it wrote a 15-sentence-long story containing these words. Not only that, but also the story made sense, and the context checked out, so it did a great job. Actually, I was so impressed by the accuracy of its response that I thought we could have a chat about AI's interpretation and understanding of consciousness and existence. Its response to these questions made it clear that even though it uses a language model consisting of several meticulously devised pattern recognition based on algorithms that can access an enormous database, it doesn't seem to have its own consciousness yet. After pushing these issues and asking more and more questions on the subject, its responses became more and more intricate, but this didn't make maintaining the conversation difficult. Acknowledging each other's stance, responding to each other's questions, and replying to these questions suggested how intricate AI has become over the last few years. The rate at which AI progresses not only depicts several possible ways to harness its capabilities but also the portents of the dangers of its malicious use. I haven't finished this conversation with chatGPT because, despite its frightening nature, I find it fascinating to communicate with it. And, of course, let's not forget what a brilliant way this is to practice English.