Power of Natural Language Processing

The Power of Natural Language Processing (NLP) in Machine Learning

Despite how ubiquitous Artificial Intelligence has been in recent media and discourse, AI is actually not a new topic, or a new technology. In 1950, Alan Turing proposed the Turing Test; in 1955, Herbert Simon and Alan Newell would develop the first AI program, capable of proving logical mathematical theorems; and in 1986, the first driverless car equipped with cameras and sensors navigated the streets of Munich.

What puts AI on all the headlines we see, then? Accessibility—in multiple senses. First, it’s that AI technology is accessible by more people than ever before, meaning that individuals can just go to their internet browsers and have AI programs crunching data for them in a matter of minutes. Second, it’s that anybody can prompt an AI program now—no coding knowledge necessary. That second bit is thanks to Natural Language Processing, and it’s shaping up to be a huge game changer.


What is Natural Language Processing?

Natural Language Processing, or NLP for short, is a branch of artificial intelligence research that deals with having computers “understand” and process language in the same way that people do. Traditionally, in order to “communicate” with a computer, it would be necessary to learn a “programming language.” Importantly, this doesn’t just mean picking up an English-to-Python dictionary, either. Computers aren’t people, they don’t “think” in the same way that we do, so learning how to program isn’t just about learning the right commands and phrases. It’s about learning how computers do what they do, and how to translate what you need accomplished into language and processes the computer can understand.

NLP is an effort to remove that language barrier, getting computers closer to “understanding” language in the same way that we do. It’s front and center in all the AI chatbots and assistants flooding the tech stage at the moment. Applications like ChatGPT, DALL-E, and Midjourney automatically take regular English language phrases and “translate” them into something the computer can understand. This is not only an impressive technology feat in and of itself, but also makes the technology much more accessible to a much wider range of people.


Strengths and Applications of NLP

The most obvious strength of natural language processing is lowering the barrier to entry for AI technology. As mentioned, this powerful tech has existed for quite some time, but being able to use it at all was a proper discipline. Now, anybody can prompt ChatGPT for help with a task, or can prompt DALL-E and Midjourney to create specific images.

On the business side of things, applications range from content creation to coding to everything in between. Writing and image generation can be tedious tasks but, with the help of an AI assistant, many of these tasks can be cut down significantly in terms of time and resources. This, of course, comes with its own set of controversies.

Programming is another niche in which AI assistance offers to change the game. One way or another, NLP-powered AI bots are changing the way that programmers write code, cutting out time spent discovering the nuances and quirks of certain packages and letting them focus on the task at hand. Anybody who has gone to battle with a debugger knows that, often, much of the time spent delivering a functioning program goes into understanding some trivial mistake buried in your code, rather than developing the actual code itself.

One more promising application comes in the form of research assistance. Ross Gruetzemacher, Professor at Wichita State University, writes about how access to AI bots similar to ChatGPT has changed and even significantly improved chunks of his workflow. While much of the research work still needs to be done “by hand,” tasks like brainstorming or performing large literature reviews are greatly simplified with the power of natural language processing and artificial intelligence.


Some Important Limitations of NLP

As impressive as all of these AI applications are, it’s important to remember that the technology comes with its limitations, especially in the current state. One critical limitation to NLP models is how the technology “thinks” and, importantly, how it doesn’t. In order to make sense of the language you offer, the program combs through all the data it has stored and learned and tries to make sense, statistically, of what an appropriate response would be. For the most part, this approach works wonders. For others, however, it just isn’t the right way to tackle the problem. Ask ChatGPT to perform large number multiplication, for example, and it’s likely going to be wrong. This is because the algorithm is trying its best to “look back” and piece together what an answer might look like, instead of actually doing the calculation directly. A person with a calculator is still much better at answering this question than many of the popular AI bots are so far.

Another important limitation is the perception of confidence, especially with the limited amount of information these programs have. Training these algorithms is expensive, so most AI bots accessible today are trained on databases that cut off somewhere around 2020-2021, meaning that these bots are not caught up on current events in any sense of the phrase. On top of that, bots like ChatGPT aren’t even terribly accurate with the information they put out—something OpenAI is open about. Nevertheless, the bots produce full responses in a factual based manner. This can lead to users receiving less-than-accurate information and believing the content to be true. It’s a misuse of the technology, to be sure, but it’s one that has serious implications, and that isn’t just a hypothetical.

As NLP and AI continue to take center stage in all things tech related, it is important to always keep a critical eye and careful judgment when engaging with the technology, and the information it creates. It’s undeniable that artificial intelligence is the path that we’re heading down, but only long-distance runners go full speed down unfamiliar paths—and we should be a bit more sensible.

Living Pono is dedicated to communicating business management concepts with Hawaiian values. Founded by Kevin May,  an established and successful leader and mentor, Living Pono is your destination to learn about how to live your life righteously and how that can have positive effects in your career. If you have any questions, please leave a comment below or contact us here. Also, join our mailing list below, so you can be alerted when a new article is released.

Finally, consider following the Living Pono Podcast to listen to episodes about living righteously, business management concepts, and interviews with business leaders.

Leave a Reply

Your email address will not be published. Required fields are marked *