By Sherman K. Newman, Contributing Writer
The past couple years have seen Artificial Intelligence become an increasingly common topic in our mainstream news. Its rapid growth and seemingly limitless potential will no doubt impact our lives in countless ways. Platforms like OpenAI, Bard, and AI Research Labs continue to create AI models that outperform humans in similar tasks. The use of AI continues to spread as new companies emerge with their own AI models, revolutionizing a growing number of industries.
Understanding AI terminology and what this new technology can do will better equip you to find employment in a rapidly evolving job market. Let’s begin by talking about how AI chatbots work.
AI would not exist without the combined technologies of natural language processing and machine learning. Machine learning is the study of computer algorithms that improve automatically through experience and has been central to AI research since the field’s inception.
“Natural languages can take different forms, such as speech, singing, or writing; a natural language or ordinary language is any language that has evolved naturally in humans through use and repetition without conscious planning or premeditation.
“Natural language processing is a field of computer science, artificial intelligence, and computational linguistics concerned with the interactions between computers and human (natural) languages and, in particular, concerned with programming computers to fruitfully process large natural language corpora,” according to @Wikipedia/Natural_Language_Processing.html.
Popular AI models, like ChatGPT and GPT-4, are fine-tuned using a process called reinforcement learning with human feedback, also known as RLHF to produce responses that are better aligned with the user’s intent. RLHF gives the model a set of questions that steer it to respond as a human would. Specifically, RLHF trains the large language model, known as LLM, to respond in the correct manner, inoffensively, ethically, and non-criminally.
GPT TECHNICAL REPORT
MARCH 2023
Users of ChatGPT and similar models make requests in the form of an instruction or question called a prompt. For example, you could ask ChatGPT to create a resume, a project proposal, or even a poem.
Where do AI models get their information? The Internet, of course. AI engineers scrape the Internet for any content — text, images, video and audio — needed to train the AI’s main component, which is the LLM.
According to the GPT-4 Technical Report:
“Large language models are being deployed in many domains of our lives ranging from browsing to voice assistants, and have potential for vast societal impacts.”
LLMs can be trained using a range of architectures and are not limited to transformer-based models. LLMs can process and produce various forms of sequential data, including assembling language, protein sequencers and chess games, extending beyond natural language applications alone.
“The role of complementary technologies remains to be seen, but maximizing the impact of LLMs appears contingent on integrating them with larger systems.”
“While the current focus is primarily on the generative capabilities of LLMs, it is important to note that these models can also be utilized for various tasks beyond text generation. For example, embedding from LLMs can be used for custom search applications, and LLMs can perform tasks such as summarization and classification where the context may be largely contained in the prompt.”
“It is essential to view LLMs as versatile building blocks for creating additional tools. Developing these tools and integrating them into systems will require time and possibly significant reconfiguration of existing processes across various industries.”
The possibility that LLMs could be classified as a general-purpose technology requires LLMs to meet three criteria: improvement over time, pervasiveness throughout the economy, and the ability to spawn complementary innovations.
LLMs on their own can have pervasive impacts across the economy, and complementary innovations enabled by LLMs — particularly via software and digital tools — can have widespread application to economic activity.”
Hopefully this article has helped to demystify some of AI’s complexity, or at least sparked an interest in wanting to learn more. We’ve only scratched the surface.
