This is a sample article I did for a company, but this exact article was never published. So, I decided to post it here as an example of my reach as a writer.
“Artificial intelligence will reach human levels by around 2029,” says Ray Kurzweil, inventor, futurist and technology writer. “Follow that out further to, say, 2045; we will have multiplied the intelligence, the human biological machine intelligence of our civilization a billion-fold.”
Many others say to prepare for the coming storm of AI to hit within the next year or two. With technology moving this fast, we need to have a basic understanding of some of its essential components and how the current state of the technology works.
Two systems creating all the buzz in the AI marketplace come with the acronyms NLP (Natural Language Processing) and NLG (Natural Language Generation). These are the principle technologies surrounding all the AI applications which are becoming seamlessly integrated into our everyday lives.
It can get very complex, but let’s explore these two different but intertwined technologies enough to put you fluid in the marketplace conversation and be able to accurately talk with customers as to what is currently happening.
What are NLP and NLG technologies?
Let’s start with NLP, the cornerstone process making the largest footprint in our lives right now. Natural Language Processing is the technology responsible when machines convert text or speech into data. We use examples of this every day if we use the speech to text on our phones, we speak to SIRI or Alexa, or use some speech to text software to create documents.
NLP now goes one step further by actually analyzing the intent of the words and can understand the intent of the content. Some developers often refer to NLU or Natural Language Understanding as the understanding portion of the NLP technology. NLU then becomes a subset of the NLP umbrella. NLU focuses on breaking down pieces of the inputted content and categorizing it into chunks to pull out the intention of the message.
Applications use Natural Language Generation (NLG) to turn text into data. Machines can now pull and present data from spreadsheets or databases and turn them into understandable reports. Consider this a digital data mining project. It used to take many labor hours to pull together reports from different databases and present them in a useable human relatable format. Now, NLP technology cuts that time down significantly.
Do NLP and NLG ever intersect?
Ideas vary about how NLP and NLG coincide. The most relatable way to understand how NLP and NLG work is to think of NLP as the overarching umbrella in which NLG and NLU exist.
Think about your experience with Alexa or Siri. SIRI takes your voice command and processes it into strings of data, the primary process behind NLP. Then it pulls apart your message and categorizes different parts into different types of data to better “understand” your command. Then the application will generate a response to the data input you requested. Siri finally either completes the command or speaks out a response.
The process of going from speech to data, understanding the input, then going from data to output could fall into all three categories.
How someone classifies the difference between NLP and NLG relies much on the type of application using the technology.
For a simple data mining system to correlate reports, the NLG technology stands on its own.
What types of technologies use NLP and NLG?
Some of the most popular uses of the NLP technology is speech recognition for text creation or software figure intent of the content. Software like Dragon Naturally Speaking simply pulls text from speech. Email anti-spam software pulls apart and analyzes content for its intention, filtering out anything which doesn’t follow a specific algorithm.
Retailers use NLG technology to create product descriptions or financial reports where tons of data needs mining. It then churns out a cohesive readable report. I’ve heard recently of some newspapers using the technology to create short automated news pieces for publications.
Could this signal an end to the writer? Probably not. Although it will likely generate less bias based material, the technology still appears a ways off from creating content with a compelling angle. But if Ray Kurzweil’s prediction comes true, we may see compelling content and attention-grabbing headlines created by AI very soon.