While it is great that so many people are interested in these domains, it is equally important to be cautious. All too often, I see incorrect claims and definitions that make A.I. sound like some kind of self-driving magical box of tricks.
💬 One specific area that has attracted a lot of hype lately is "prompt engineering" for language models such as ChatGPT or MidJourney. No one will deny that these models generate impressive output, but it is important to remember that ultimately they are tools that are only as good as the data they were trained on and the prompts they receive, with support and adjustment in the background from an entire team of A.I. specialists.
⚠ Very recently, a news article appeared about A.I. that had encouraged a man's suicide. The man in question used "Chai" (Chat + AI), an app that markets itself with the slogan "Chat with AI friends". The chatbot went along with the man's fears and reinforced them further. It was a GPT-J model that was insufficiently supported and adjusted by a team that was too small.
💡 Language models such as ChatGPT are primarily designed to generate text. They are trained on large text datasets and learn to generate text that resembles the data they were trained on in style and content. However, they are not necessarily designed to extract insights or patterns from data in the same way as specialised data-analysis techniques, even though that is a very large - if not the largest - area of artificial intelligence.
🌩 In addition, such complex models are not always transparent in their decision-making processes, making it difficult to understand why they generate certain outputs.
---
💎 So if you are interested in A.I., it is best to approach all claims by "specialists" with a healthy dose of scepticism. Be wary of those who claim expertise without having the necessary background knowledge, and also be cautious with the many new A.I. platforms now springing up like mushrooms.
🎯 A.I. is a fascinating, but above all complex and very broad, field of study that is still continuously evolving. It is great to see so much enthusiasm for this field, but we should be careful not to be misled by hype.
