Today’s news is filled with articles and conversations about a new technology called ChatGPT. What is it and why is it in the news?
The world of technology is filled with acronyms and phraseology that is often comprehendible by only those who work within the industry. Usually this is nothing more than shorthand or abbreviations that make conversations between industry professionals go faster, if you speak the language. It is no different than the medical or legal professions with their own terms and jargon. The media has a knack for picking all this up and using it in articles as if they understood it.
What makes this program (ChatGPT, GPT-3, and GPT-4) a bit different is that companies start with software that is “open source,” meaning available to anyone for free or at a nominal fee. They then use this as the starting point on which other, more sophisticated programs are built. Using ChatGPT, major technical companies can build interactive software that performs like virtual assistants, but do this in a meaningful, conversational way.
In the most visible case Microsoft recently announced that the search functions for Bing which runs on their internet browser, Edge, will now use ChatGPT as the way it will provide search answers. The software is already being used and becoming widely accepted, quickly. With GPT-4 the scenario is a little different because the starting point is more sophisticated than many other programs when they finish.
ChatGPT was built by the company OpenAI, which is governed by a non-profit. Their stated mission is:
“…to ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity.”
And they have a very noble set of principles abbreviated as:
“Broadly Distributed Benefits – We commit to use any influence we obtain over AGI’s deployment to ensure it is used for the benefit of all, and to avoid enabling uses of AI or AGI that harm humanity or unduly concentrate power…”
“Long-term safety – We are committed to doing the research required to make AGI safe, and to driving the broad adoption of such research across the AI community…”
“Technical leadership – To be effective at addressing AGI’s impact on society, OpenAI must be on the cutting edge of AI capabilities—policy and safety advocacy alone would be insufficient…”
“Cooperative orientation – We will actively cooperate with other research and policy institutions; we seek to create a global community working together to address AGI’s global challenges…”
These are lofty missions and principles and whether OpenAI can live up to them is the key not just to the success of GPT but to our world. The potential of this software is so great that it is not OpenAI but those who will build other programs from it that are the challenge. OpenAI is admitting in its goals that they want to be sure the technology is not misused, acknowledging that there is great potential for abuse.
It seems to us that these lofty goals will be difficult to administer, but at least someone sees it as their mission to try. In just the first two months of availability ChatGPT acquired 100 million users, the fastest adoption of any new technology in history, and this is likely an understatement of the users.
What is ChatGPT and Its Other Iterations?
We need to start by breaking down the acronym into its component pieces.
Chat
Chat refers to nothing more than the ability of this software to listen or read normal conversations and interpret that input into useful information. It is the starting point for anyone to interact with the program. Nothing more than ask a question to get an answer in normal conversational terms. This process also labels the program as a chatbot.
GPT stands for Generative Pretrained Transformer. Let us break this down into these three components.
Generative – Generative is a key term and function of ChatGPT. It refers to the software’s ability to recognize patterns or traits within existing data, and then create new content based on the analytics. For example, the query “What are the three best operas of all time?” Using a normal search engine will give you a list of web sites to visit to get the answer. ChatGPT responds with a list of the ten best operas of all time, and then gives you fifteen references to show how it arrived at the conclusion. It gives you the answer, not the interim step we are accustomed to with simple search engines.
Pretrained – Pretrained means nothing more than that. The software comes with several functions such as speech recognition, pattern recognition, and industry specific training. A good example would be adding the knowledge of street signs as a starting point for automobile manufacturers so that they do not need to train the software on their own. In this way street sign recognition is a basic function of ChatGPT and the car manufacturer does not need to create this capability. To create this type of functionality though, ChatGPT must be exposed to billions, even trillions of images, documents, and other points of contact to create the knowledge.
Transformer – Transformer refers to the capability of the software to receive vast quantities of traditional information and “transform” it into a machine format that can then be used for comparison to its pre-existing information. It differs from earlier software in its ability to do “deep learning” for functions like natural language recognition or visual recognition of images or art.
OpenAI is working on many other aspects of artificial intelligence, and these will only enhance the software’s ability to function in a more human-like manner.
Let us summarize what we believe a good, straightforward definition of ChatGPT is for non-technical people. ChatGPT is a set of computer programs with the ability to receive requests for information in a normal, conversational format of either written or spoken questions. The program then converts that request into a form that can be used to analyze stored information in the pursuit of an answer. Additionally, ChatGPT has the ability to store, analyze, and index vast quantities of information used to produce these answers. Within the ability to produce answers to questions, ChatGPT can combine its analysis into unique or creative responses.
ChatGPT and other competing AI products have already demonstrated a capability to write compositions, compose music, and take tests like the SAT and LSAT with at least 80% competence. It can also create art based on descriptions of desired results.
Artificial Intelligence capabilities are already here, and they will be coming at greater speeds than we have seen with any other technology. The potential for humanity is enormous and great leaps forward in different segments of our world will be possible. ChatGPT is the most visible because of the Microsoft interface through Bing, but it is only one of many AI programs headed our way. Fearing a major loss of search revenue, Google responded with its own test version of ChatGPT this week. It is a test and does not work well, but it shows the pressure one company can place on another with just a slight lead in technology.
With great capability comes great responsibility, and great opportunity for abuse. We will tackle those issues in the next installments.
Resources Used In These Articles
36 Artificial Intelligence Examples Shaking Up Business Across Industries, By Sam Daley, updated by Parul Pandey and Matthew Urwin, BuiltIn.com, February 17, 2023.
Advantages and Disadvantages of Artificial Intelligence, By Nikita Duggal, simplilearn.com, March 9, 2023.
All the Benefits of Artificial Intelligence, Western Governors University, wgu.edu, April 25, 2022.
Big Data Statistics 2023: How Much Data is in The World?, by Ogi Djuraskovic, FirstSiteGuide.com, December 26, 2022.
ChatGPT explained: everything you need to know about the AI chatbot, By Mark Wilson, TechRadar.com, March 15, 2023.
Don’t overlook independence in Responsible AI, by Editorial Team, insideBIGDATA.com, March 17, 2023.
How Fast Is Technology Advancing in 2023?, by Jacquelyn Bulao, techjury.com, February 27, 2023.
How Fast Is Technology Growing – Can Moore’s Law Still Explain the Progress?, By Nick Gavon, WebTribunal.com, March 4, 2023.
How Fast Is Technology Growing Statistics [Updated 2023], By Darina L., March 7, 2023.
How Much Data Is Created Every Day? +27 Staggering Stats, by Branka Vuleto, SeedScientific.com, October 28, 2021.
OpenAI Charter: Our Charter describes the principles we use to execute on, OpenAI’s mission, OpenAI.com, last accessed March 19, 2023.
Pros and cons of AI: Is Artificial Intelligence suitable for you?, By Eray Eliacik, Dataconomy.com, August 19, 2022.
Sparks of Artificial General Intelligence: Early experiments with GPT-4, By Microsoft Research, ~March 24, 2023.
Sparks of Consciousness: The Era Of Artificial General Intelligence, By Scott Porter, LinkedIn.com, March 27, 2023.
The Exponential Growth of Data, by Editorial Team, insideBIGDATA.com, February 16, 2017.
Top 7 Different Uses of Artificial Intelligence, by Java Assignment Help, JavaAssignmentHelp.com, December 15, 2021.
Transformer Neural Networks, DeepAI.com, March 20, 2022.
Unintended Consequences, by Rob Norton, UPenn.edu, Date Not Cited.
What’s the real story behind the explosive growth of data?, by Ulrike Hack, redgate.com, September 8, 2021.
Why Elon Musk and Steve Wozniak have said AI can ‘pose profound risks to society and humanity,’ by Explained Desk, The Indian Express, indianexpress.com, April 3, 2023.
Xi, Putin declare intent to rule the world of AI, infosec, by Simon Sharwood, TheRegister.com, March 22, 2023.
