top of page
Dale C. Maley

Will AI revolutionize our lives?





A new Artificial Intelligence (AI) software program called ChatGPT was released to the public for initial testing in November 2022.


After a year and a half of testing, the production version of ChatGPT was released in August 2024. This new AI program has generated a host of other AI programs that have excited the public, businesses, and investors. This new software offers the potential to revolutionize many fields, including historical research.

 

Most people assume that AI is a new phenomenon that has only occurred in the last couple of years. In reality, AI dates back to the 1950s. Computers and their associated software received a lot of attention during World War II because they could speed up things like creating tables for how to aim a weapon and hit a target far away.

 

Credit for the words Artificial Intelligence is given to Marvin Minsky and John McCarthy, who hosted the first workshop on the topic at Dartmouth College in 1956. Following this workshop, AI experienced its first "boom" period, resulting in two decades of success in funding and technological progress.

 

AI research and results have been a cyclic phenomenon since the 1950s. Veterans of this industry refer to the boom and bust periods as AI Summers and AI Winters. The two decades after the 1956 Dartmouth conference are referred to as the first Summer phase.

 

Due to a lack of computing power, interest in AI declined, and the industry entered its first Winter period in the late 1970s.

 

The 1980s saw the second Summer phase of AI with the promise of developing Expert Systems that could replace human judgment. The general idea was that Expert Systems could learn the rules of logic that human experts used to solve problems. Once the Expert System learned the right rules, it could solve problems better and faster than human beings. For several reasons, Expert System programs could never deliver the improvements initially forecasted, and the industry entered another Winter period.

 

Around 2010, the AI industry entered its third boom period, and the phrase "Machine Learning" was created. Each year, the amount of new information added to the Internet grows exponentially, and computing power has also grown exponentially.

 

A new concept was developed that could use massive computers to store all the accumulated knowledge on the Internet. Once all of this knowledge was captured, it could be used to solve problems quickly and efficiently. The initial use of ChatGPT was to create new documents such as essays, biographies, obituaries, and stories.

 

The traditional method of researching and writing local history stories has been for a human to search the Internet for all the information available about a topic. This process is very labor intensive, and many different sources on the Internet must be individually searched. An experienced historian then collects and organizes all of this information into an exciting and coherent history story.

 

While ChatGPT was still in its initial testing phase, an experiment was conducted. First, conventional research methods were used to learn everything possible about John Paul Yost from Pontiac, Illinois. This information was used to help write the history book The Zoath Yost Family of Pontiac, Illinois, by Dale C. Maley.

 

ChatGPT was then asked to write a biography about John Paul Yost from Pontiac, Illinois. The ChatGPT program responded with a well-written biography about Mr. Yost, with nine different facts about his life. The AI-generated biography read well and sounded impressive. ChatGPT did not cite any references for its nine facts.

 

Unfortunately, every single one of the nine facts was wrong! When ChatGPT was asked for its reference source for the first fact, the AI program apologized and removed that fact from the biography. Each successive time that ChatGPT was asked for the reference where it found the fact, it said there was no reference and removed the fact from the biography.

 

Although ChatGPT did not give accurate answers to questions, it became extremely popular with the general public because it does perform some tasks very well. The popularity of ChatGPT spawned the creation of many different types of AI programs.

 

Another type of AI program is Perplexity. It was first released in early 2023 and cites a reference for every fact it includes in its responses. In October 2024, the Perplexity AI program was asked to write a biography for John Paul Yost of Pontiac. Perplexity's resultant biography was very detailed and very accurate. The primary source for all of its facts was the recently published history book The Zoath Yost Family of Pontiac, Illinois, by Dale C. Maley.

 

Other users of AI programs have reported great success in using some of the new AI programs to save time on common tasks. Some of these everyday tasks include creating new posts on Facebook, writing papers, creating YouTube videos, and creating PowerPoint presentations. Some new AI programs allow the content creator to use their own voice to convert written text to speech for videos.

 

A recent paper by anonymous authors about the AI industry is titled "Is the Next Winter Coming for AI?" In this paper, the authors created an interesting chart documenting the boom and bust periods of AI over the last 70 years. The AI industry is currently in a boom or Summer period in its cycle.

 

The interesting question is what direction the AI industry will take in the future. The authors note that one option is that current momentum stalls out, and the industry goes into another Winter period. A second option is that AI becomes pervasive and is used in many different applications. The third option is that current momentum tapers off, but the AI industry remains relevant to the general public.

 

Based on the examples of ChatGPT and Perplexity noted above, AI momentum will reduce slightly, but the AI industry will remain an important productivity tool for users.

 

It will be interesting to see if AI remains popular or enters its third Winter period over the next decade. Current users should always review and question the accuracy of any information generated by today's AI programs.


(Dale Maley's weekly history feature on Fairbury News is brought to you by Dr. Charlene Aaron)

 

 

110 views0 comments

Recent Posts

See All

Comentários


DAVES LOGO larger.jpg
Image.jpeg
bottom of page