Thread by @socworkpodcast

One of the hot topics at #SSWR2023 was #ChatGPT and AI chat. I wanted to provide a short primer for #SocialWork educators and practitioners about what's going on behind the scenes. I'll cite my sources for this thread at the end. 🧵

#ChatGPT is a chatbot that uses @OpenAI's large language model (LLM) GPT-3.5 to generate long strings of text that resemble content produced by human writers. LLM is a type of artificial intelligence that uses deep learning to understand and generate natural language. /1

Deep learning is a type of machine learning that uses artificial neural networks with multiple layers to learn and make decisions. In LLM, deep learning is used to train the model to understand and generate natural language tasks (e.g. language translation, text summarization) /2

The model learns to recognize patterns and relationships in the language, allowing it to generate human-like text. The use of multiple layers in the neural network allows the model to learn increasingly complex representations of the data, making it more accurate and powerful. /3

This is why deep learning is considered a key component in the development of advanced language models like GPT-3. LLMs are trained on massive amounts of text data (sometimes called "bags of words"). /4

While LLMs have the potential to revolutionize industries such as social work, customer service, content creation, and language education, there are also concerns about their potential misuse and impact on jobs. /5

As with all aspect of #SWTech, it's important to consider the ethical implications and potential consequences of using LLMs in various applications, and to continue to used them in practice, research and education responsibly.
cc: @igraywill @JimmySW @ChenglinHong
/6

Source: I combined the results of three @OpenAI ChatGPT requests to explain LLM and deep learning. I did some slight wordsmithing to stitch the results together. Images created in @midjourney. Each image has the specific prompt as the "alt" text. /end