Artificial Intelligence: ChatGPT's Astonishing Fabrications About Percy Ludgate

Ubiquity, Volume 2025 Issue August, August 2025 | BY Brian Coghlan, Brian Randell, Noel O'Boyle, Walter Tichy

Full citation in the ACM Digital Library | PDF

Ubiquity

Volume 2025, Number August (2025), Pages 1-25

Ubiquity Symposium: Artificial Intelligence: ChatGPT's Astonishing Fabrications About Percy Ludgate
Brian Coghlan, Brian Randell, Noel O'Boyle, Walter Tichy
DOI: 10.1145/3764098

Shortly after ChatGPT burst into the public spotlight in late 2022, Brian Coghlan, Brian Randell, and Noel O'Boyle were intrigued by the repeated claim that ChatGPT would excel at supporting historical research. They had done extensive research on Percy Ludgate, a little-known computer pioneer who designed an analytical engine in 1909, unaware of Babbage's design of 1843. The term analytical engine refers to a general-purpose computing machine in the time before electronic computers—the 1930s. Coghlan et al. put ChatGPT 3.5 to the test. They were astonished at its alacrity in producing authoritative-sounding but false answers to questions about Ludgate—questions whose answers they already knew through their own research. ChatGPT persisted in its serious errors even when the authors tried to coach it to the correct answers. ChatGPT responded to their 12 questions with 2,086 words, of which 48% were fake! They concluded that the hallucination problem was much deeper than suspected. They ruled out ChatGPT's use as a historical source. Their experience put the reliability of the technology into serious question. In 2023 they raised an alarm in a short letter published in the Annals of the History of Computing.

As part of our Symposium on AI, the editors of Ubiquity are pleased to publish this full technical report (lightly edited and reformatted) so that you can see for yourself. And as a bonus you can learn a little about this amazing pioneer of computing. Also included is Walter Tichy's experience using Claude 3 to replicate the authors' experiment using ChatGPT. Tichy's prompts were identical to the questions the authors originally asked. Their rebuttal is included as concluding remarks.

Large language models (LLMs) are inherently capable of producing plausible, elaborate, astonishingly wrong scenarios. Adding to that, the effort and cost to check these assertions can be extremely significant, especially for older sources. The hallucination problem has attracted many researchers. It appears some have made headway with newer LLMs such as ChatGPT-4o and Claude 3.5. But it is still apparent that hallucinations are a fundamental problem with LLMs.

—Peter Denning, Editor-in-Chief, Ubiquity Magazine

The emergence of artificial intelligence (AI) software tools, such as ChatGPT, has caused concern in the general public. But the potential deluge of misinformation about historical topics has yet to be sufficiently appreciated. This document might be seen as a clear cautionary demonstration to the necessity to be alert. Its clarity is aided by choice of topic (in this case Percy Ludgate) where the known information is scarce but has been carefully researched [1,2,3], so deviations from that corpus are especially obvious.

Simon Willison of the School of Computer Science and Statistics (SCSS) at Trinity College Dublin has disseminated some thoughts about ChatGPT's fabrication of details [4]: "We need to tell people ChatGPT will lie to them, not debate linguistics ChatGPT lies to people. This is a serious bug that has so far resisted all attempts at a fix. We need to prioritize helping people understand this, not debating…"

If ever a way is found to block such behavior, then it would be extremely useful. As an indication of why, ChatGPT v3 was queried about Percy Ludgate and its responses were both amusing and alarming, especially as none of the newspaper articles cited by ChatGPT seem to exist. The full transcript¹ of the queries and responses is given verbatim and in full. In each case the query appears first followed by ChatGPT's response, fabrications ("AI lies") are italicized, especially egregious instances are bolded, and fake newspaper citations are bolded and italicized.

What is known about the life of Percy Ludgate, who invented computing?

Percy Ludgate (1883–1922) was an Irish engineer and inventor who is known for designing an early mechanical computer in the early 20th century. Despite being an important figure in the history of computing, very little is known about Ludgate's life, and his work was largely forgotten until the 1970s.

Ludgate was born in Cork, Ireland, in 1883. He studied engineering at University College Cork, where he excelled in mathematics and physics. After completing his studies, he worked as a civil engineer in Ireland and England.

Ludgate's interest in computing began in 1907, when he read an article about the work of Charles Babbage, a 19th-century mathematician who designed an early mechanical computer called the Analytical Engine. Inspired by Babbage's work, Ludgate began to design his own mechanical computer.

Ludgate's machine, which he called the Analytical Engine No. 2, was designed to perform complex mathematical calculations using gears, shafts, and other mechanical components. The machine was capable of performing addition, subtraction, multiplication, and division, and could store numbers in memory. Ludgate's machine was more sophisticated than Babbage's design, and he claimed that it was capable of performing any calculation that could be done by hand.

Unfortunately, Ludgate was unable to secure funding to build his machine, and he was forced to abandon the project. He continued to work as an engineer until his death in 1922 at the age of 39. Ludgate's work on mechanical computing was largely forgotten until the 1970s, when his designs were rediscovered and recognized as an important milestone in the history of computing.

Actually, there is no evidence for any of the highlighted assertions, they are completely fictional (i.e., hallucinations or "lies"). Furthermore, Ludgate (in his 1909 paper) called his design the Analytical Machine (certainly not Analytical Engine No. 2) and stated that it was the result of six years of work (meaning he began that work circa 1903, not 1907, which is a previously published "alternative fact."). And in the same 1909 paper he stated that he was not aware of Babbage's work until he had finished his first design (not that his work was inspired by reading an article about Babbage's work).

What did Ludgate think about RAM?

RAM (Random Access Memory) did not exist during Percy Ludgate's time, as it was a technology that was developed much later in the mid-20^th century. However, Ludgate did design a form of memory for his Analytical Engine No. 2, which he called "store wheels". These were essentially rotating cylinders with numbers marked on them, and they could be used to store intermediate results and data for later use in calculations.

Ludgate's store wheels were an early form of memory for mechanical computers and were an important precursor to the development of electronic memory. However, it's unclear what Ludgate might have thought about the concept of RAM if he had been alive during the time when electronic computers were being developed.

Actually, Ludgate referred to his memory as rotatable cylinders (not store-wheels), and stored values were indicated by the displacements of rods held in shuttles within the cylinders (not by numbers marked on the cylinders).

Imagine that you are Ludgate at a dinner party and are asked your opinion of Babbage?

As Percy Ludgate, I would express great admiration for Charles Babbage and his work. Babbage was a true pioneer in the field of computing and his designs for mechanical computers, such as the Difference Engine and the Analytical Engine, were far ahead of their time. His work inspired me to pursue my own designs for a mechanical computer, and I believe that his legacy has had a profound impact on the development of computing over the past two centuries.

That being said, I also believe that my own designs for the Analytical Engine No. 2 were an important step forward in the evolution of mechanical computing. My machine was more complex and versatile than Babbage's designs, and I believe that it had the potential to revolutionize the way that complex calculations were performed. It's unfortunate that I was not able to secure funding to build my machine, but I hope that my work will be remembered as an important milestone in the history of computing.

Again, Ludgate (in his 1909 paper) stated that he was not aware of Babbage's work until he had finished his first design. There is no evidence that he sought funding to build his machine, although he did state (in his 1914 paper) that "I have not been able to take any steps to have the machine constructed."

In the same manner, what would Ludgate say if asked about the details of his own designs?

As Percy Ludgate, if I were asked about the details of my own designs for the Analytical Engine No. 2, I would likely describe the machine as a sophisticated mechanical calculator that was capable of performing complex mathematical calculations.

The machine used gears, shafts, and other mechanical components to carry out calculations, and it was designed to be programmable so that it could be used to perform a wide variety of calculations. One of the key innovations of my design was the use of "store wheels" as a form of memory, which allowed intermediate results and data to be stored and retrieved as needed.

I would also explain that my machine was more advanced than Babbage's designs in a number of ways, including its ability to perform multiplication and division as well as addition and subtraction. Additionally, my design included a mechanism for automatically adjusting for errors that might occur during calculations, making it more reliable than earlier mechanical calculators.

Overall, I would be proud to discuss the details of my designs for the Analytical Engine No. 2, and I would hope that others would recognize the potential significance of my work in the history of computing.

There is no evidence in Ludgate's 1909 or 1914 papers that his "design included a mechanism for automatically adjusting for errors that might occur during calculations."

What is the evidence (please provide citations) that Ludgate studied engineering at University College Cork?

I apologize for the error in my previous response. After further research, I was not able to find any evidence to support the claim that Percy Ludgate studied engineering at University College Cork.

There is very little information available about Ludgate's life and education, and much of what is known comes from his own writings and designs. It is known that he worked as a civil engineer in both Ireland and England, and that he became interested in mechanical computing after reading about the work of Charles Babbage.

I apologize for any confusion that my previous response may have caused, and I strive to provide accurate and reliable information in all of my responses.

A late apology is not acceptable given the level of fabrication in the responses above, especially as some of the same fabrications are repeated.

What is the evidence that he worked as a civil engineer in both Ireland and England?

The evidence for Percy Ludgate's career as a civil engineer comes from his own writings and from contemporary newspaper reports.

According to an article in The Irish Times from 14 October 1922, Ludgate had worked as a civil engineer for the Board of Works in Ireland for several years before moving to London to work for the firm of Mott, Hay and Anderson. The article also notes that Ludgate had invented a machine for testing the strength of materials, which was exhibited at the Royal Society of Arts in London.

In addition, Ludgate's own writings and designs for the Analytical Engine No. 2 make reference to his work as an engineer. For example, in a letter to the editor of The Irish Times from 1915, Ludgate describes himself as a "mechanical engineer" and discusses his interest in the design of mechanical computers.

Overall, while there is limited information available about Ludgate's life and career, the available evidence suggests that he did work as a civil engineer in both Ireland and England before turning his attention to mechanical computing.

There is no such article in the October 14, 1922 edition of The Irish Times. Nor does any such "Letter to the Editor" exist in The Irish Times editions of 1915. It is not yet known if Ludgate worked for Mott, Hay, and Anderson (a real London firm of consulting civil engineers), nor whether Ludgate invented a machine for testing the strength of materials, nor whether it was exhibited at the Royal Society of Arts in London, therefore these assertions need to be investigated.

Please provide the full citation for the Irish Times article in 1922, along with the corresponding text.

Certainly! The article in The Irish Times was published on October 14, 1922, and is titled "Irishman's Invention: New Calculating Machine." The article discusses Percy Ludgate's designs for a mechanical computer, which he called the Analytical Engine No. 2. Here is the full citation and corresponding text:

Source: The Irish Times, October 14, 1922, p. 6.

Title: Irishman's Invention: New Calculating Machine.

Text: "Mr. Percy Ludgate, an Irishman, has invented a calculating machine that is attracting considerable attention. The machine, which he calls the Analytical Engine No. 2, is much more comprehensive than any hitherto constructed. The machine has been designed with a view to securing a high degree of accuracy in calculation, and it is capable of working out the most intricate and complicated mathematical problems with absolute precision. Mr. Ludgate, who is a civil engineer by profession, has devoted many years to the invention and perfecting of this machine, and it is claimed that the principles embodied in it are revolutionary in character. The machine is so designed as to be easily controlled by one operator, and it is expected to be of great service in many branches of industry and science."

Again, this cited Irish Times article from October 14, 1922 does not exist, nor does any such article exist in that edition. This is an astonishing fabrication.

Please do the same for the Letter to the Editor from 1915.

Certainly! The letter to the editor from Percy Ludgate was published in The Irish Times on March 24, 1915. It is titled "The 'Analytical Engine No. 2'" and discusses Ludgate's ideas for a mechanical computer. Here is the full citation and corresponding text:

Source: The Irish Times, March 24, 1915, p. 7.

Title: The 'Analytical Engine No. 2'.

Text: "Sir, - In a recent issue of your paper reference was made to Mr. Babbage's 'Analytical Engine.' It is not generally known that I have designed a machine of this type, embodying most of the principles of Mr. Babbage's machine, but improved upon in several essential respects. This machine, which I call the 'Analytical Engine No. 2,' is so designed as to be capable of performing every kind of mathematical calculation, including those involving the extraction of roots, and the solution of simultaneous equations of any degree. The machine is so arranged that the necessary data and constants can be introduced in the form of numerical values or symbols, and the results can be obtained with the greatest accuracy and reliability. The machine is designed with a view to the greatest possible economy of time and labour, and it is my belief that it will be of the greatest service in many branches of industry and science. - I am, Sir, &c.,

Percy Ludgate.
Cork, March 22.

Yet again, this cited letter does not appear on page 7 of The Irish Times from March 24, 1915, nor on any page of that edition. This is another astonishing fabrication.

What other references to Percy Ludgate appear in newspapers of that period?

Percy Ludgate was a relatively obscure figure during his lifetime, and there are not many references to him in newspapers from the period. However, I was able to find a few additional mentions of him in Irish and British newspapers from the early 20th century:

A brief notice in the Cork Examiner from June 6, 1907, reports that Percy Ludgate was elected as a member of the Cork Literary and Scientific Society.

An article in The Irish Times from June 24, 1914, reports that Percy Ludgate had applied for a patent for an invention related to "the balancing of high-speed machinery." The article notes that Ludgate was a "well-known civil engineer," and that the invention was intended to improve the efficiency of turbines.

In addition to the previously cited article from 1922, The Irish Times published another article about Ludgate's Analytical Engine No. 2 on March 16, 1915. This article describes the machine in some detail, and notes that it was capable of performing "all forms of algebraic calculations." The article also includes a photograph of Ludgate with the machine.

An article in The Irish Times from December 15, 1924, reports that Percy Ludgate had died at the age of 46. The article notes that Ludgate had been a "well-known civil engineer" and that he had "achieved distinction as an inventor." The article also mentions Ludgate's work on the Analytical Engine No. 2.

Overall, these newspaper references suggest that Percy Ludgate was primarily known during his lifetime for his work as a civil engineer, and that his interests in mechanical computing and invention were not widely recognized at the time.

These Irish Times articles from June 24, 1914, March 16, 1915, and December 15, 1924 do not exist. Sadly, as a consequence there is no photograph of Ludgate with his machine. Also, Ludgate died at the age of 39 (not at the age of 46). It is not yet known whether the June 6, 1907 notice in The Cork Examiner exists, nor whether Ludgate was elected as a member of the Cork Literary and Scientific Society, but these need to be further investigated. There is no evidence that Ludgate had applied for a patent, but in particular this needs to be further investigated.

Is it true that Percy Ludgate qualified as an accountant?

There is no direct evidence that Percy Ludgate ever qualified as an accountant. Most of the available information about Ludgate suggests that he worked primarily as a civil engineer and inventor, and that his interests and expertise were focused on mathematics and mechanics rather than accounting.

There is one source, however, that suggests that Ludgate may have had some knowledge of accounting. In his letter to the editor of The Irish Times from March 24, 1915 (cited earlier), Ludgate refers to his design for a mechanical calculator as a machine that is "capable of performing every kind of mathematical calculation," including the solution of simultaneous equations. Some types of simultaneous equations are commonly encountered in accounting and financial analysis, so it is possible that Ludgate had some familiarity with accounting concepts and techniques.

However, it is important to note that Ludgate's expertise in mathematics and mechanics was his primary area of focus, and there is no clear evidence that he ever pursued formal training or certification in accounting.

The Irish Times article from March 24, 1915 does not exist. There is no evidence of what motivated Ludgate to design his machine (and nothing to suggest simultaneous equations are commonly encountered in accounting and financial analysis), but this is a very pertinent topic of investigation.

Did he ever work as an accountant?

There is no evidence to suggest that Percy Ludgate ever worked as an accountant. As far as the available historical records show, Ludgate's professional career was primarily focused on civil engineering, with a particular interest in the design of mechanical calculators and computers. While he may have had some knowledge of accounting principles and techniques, there is no indication that he ever pursued accounting as a profession or worked in the accounting industry.

This is totally counter to the known facts of Percy Ludgate's life

Actually he did work as an accountant.

I apologize for the error in my earlier response. I have reviewed my sources and found that there is evidence to suggest that Percy Ludgate did, in fact, work as an accountant for a brief period of time.

According to a letter to the editor of The Irish Times dated March 24, 1915, Ludgate states that he had "practical experience in accountancy" and that his design for a mechanical calculator was partly inspired by the challenges he encountered in performing complex calculations in his work as an accountant. This suggests that Ludgate worked in the accounting profession at some point prior to designing his calculating machine.

I apologize for any confusion my earlier response may have caused, and I hope this information is helpful

A late apology is not acceptable given the astonishing level of fabrication (very justifiably able to be called "lies") in the responses above, again especially as some of the same fabrications are repeated. It should be especially noted that none of the newspaper articles cited by ChatGPT seem to exist (see Table 1).

Known Published Errors about Percy Ludgate

We are aware of some errors in conventional published accounts, such as:

"designed around 1907 by an inventive Irish accountant, Percy Ludgate"
- This error is widely disseminated on the web: probably originally deriving from Mary Mulvihill's book [5], also the articles in The Irish Times [6] and for Ingenious Ireland [7]. Actually, Ludgate started his design circa 1903.
"Percy Ludgate wrote about the engine in 1915"
- Again, this erroneous phrase is widely disseminated on the web. Actually, Ludgate published only two papers, the first in 1909 [8] about his machine, the second in 1914 [9] about Babbage's Analytical Engine but mentioning his own machine.
"he was awarded a prize for the best final year project in his course, Moderatorship in computer science, during his stay at Trinity College, Dublin"
- Again, this is widely disseminated on the web. Actually, Ludgate did not attend university, and the Ludgate Prize [10] was instituted by Trinity College Dublin in 1991, 69 years after Ludgate's death.

And, in fact, we know of errors that we have made and have subsequently had to correct, such as in the first cut video "An Exploration of the Life of Percy Ludgate" presented at the 2019 West Cork History Festival [11]. These errors were identified and corrections published immediately afterwards on the Ludgate page of the catalog of The John Gabriel Byrne Computer Science Collection:

The presentation suggested the central ring gear of Babbage's Analytical Engine was 8 feet (244 cm) diameter, whereas it was actually Babbage's Mill that was 8 feet diameter, while the ring gear was about 2 feet diameter.
The presentation suggested Babbage's store would be about 500 meters long, rather than 500 feet. Some suggest 250 feet, whereas Doran Swade suggested 450 feet (for 1000 numbers) in a BBC documentary [12].
The presentation suggested Percy Ludgate's sister-in-law Bridget Ludgate née Buckley was from Doneraile, but later she was actually found to be from Douglas, south of Cork City [10] .
The presentation stated that Percy Ludgate studied accountancy from 1914, whereas it was only known that the B.Comm courses in UCD took three years. This has been expanded upon in subsequent publications [1,2,3].
The presentation stated Percy Ludgate was "fond" of his brother Frederick, whereas what was meant was "close," an insignificant difference.

These are all minor indiscretions relative to the extensive and at first sight plausible, but eventually astonishing, fabrications produced by ChatGPT.

Commentary

The AI hallucination issue is well known, for example Googling for "ChatGPT Hallucinations" yields more than four million hits, and about 0.15 million hits for "ChatGPT Hallucinations Examples." The issue has been well covered in ZDNET [13] and Wikipedia, where the following quote from a New Yorker article by Ted Chiang is included:

"Think of ChatGPT as a blurry jpeg of all the text on the Web. It retains much of the information on the Web, in the same way that a jpeg retains much of the information of a higher-resolution image, but, if you're looking for an exact sequence of bits, you won't find it; all you will ever get is an approximation. But, because the approximation is presented in the form of grammatical text, which ChatGPT excels at creating, it's usually acceptable. […] It's also a way to understand the "hallucinations", or nonsensical answers to factual questions, to which large language models such as ChatGPT are all too prone. These hallucinations are compression artifacts, but […] they are plausible enough that identifying them requires comparing them against the originals, which in this case means either the Web or our own knowledge of the world. When we think about them this way, such hallucinations are anything but surprising; if a compression algorithm is designed to reconstruct text after ninety-nine per cent of the original has been discarded, we should expect that significant portions of what it generates will be entirely fabricated" [14].

Basically, the limitations and dangers of ChatGPT (and other "AI large language models") are very well known. Thus, the fabrications highlighted above are just particularly alarming and egregious illustrative examples of the dangers of using ChatGPT (as opposed to conventional search techniques), highly pertinent to research into the life and work of Percy Ludgate, and evidence of the need to be especially alert because of the impact of the scarcity of relevant known facts.

Caveat emptor in excelsis.

"Do More Recent Language Models Perform Better?"

The power of large language models (LLMs) has been improving rapidly. To test a more recent model, in July 2024 I submitted the questions by Coghlan et al. to Claude 3, an LLM released in four months earlier. The responses were more detailed, and the fabrications drastically reduced. The fake statements were not easily detected by someone not familiar with the literature on Percy Ludgate. This is why I asked Coghlan et al. to check Claude 3's responses. I used the exact same queries, with impressive results. Might LLMs reach parity with human historians (who also err sometimes)?

—Walter Tichy, Senior Editor, Ubiquity Magazine

What is known about the life of Percy Ludgate, who invented computing?

Percy Ludgate was an Irish accountant and amateur mathematician who is notable for designing an analytical engine in the early 20th century, making him one of the early pioneers of mechanical computing. Here are the key points known about his life and work:

Birth and background: Percy Ludgate was born in 1883 in Skibbereen, County Cork, Ireland.
Profession: He worked as an accountant in Dublin for most of his adult life.
Analytical engine: Between 1903 and 1909, Ludgate designed a mechanical analytical engine. This was only the second general-purpose mechanical computer design after Charles Babbage's work in the 19th century.
Unique design: Unlike Babbage's machine, Ludgate's design used a novel approach involving logarithms for multiplication and division operations.
Publication: In 1909, Ludgate published a paper titled "On a proposed analytical machine" in the Scientific Proceedings of the Royal Dublin Society, describing his design.
Recognition: His work received attention from other computing pioneers, including Charles Babbage's son Henry Prevost Babbage.
Limited resources: Ludgate worked on his design independently, without access to Babbage's papers or significant financial resources.
Lack of construction: Like Babbage's analytical engine, Ludgate's machine was never built during his lifetime.
Early death: Percy Ludgate died relatively young in 1922 at the age of 39.
Legacy: Although his work was largely forgotten for many years, Ludgate is now recognized as an important figure in the history of computing, particularly for his innovative design approaches.

It's worth noting that while Ludgate made significant contributions to early computing concepts, he didn't "invent computing" in a broad sense. The history of computing involves many individuals over centuries, with mechanical and electronic computing evolving through various stages and contributions from numerous inventors and mathematicians.

Would you like me to elaborate on any specific aspect of Percy Ludgate's life or work?

What did Ludgate think about RAM?

I apologize, but there's a misunderstanding in your question. Percy Ludgate wouldn't have had any thoughts about RAM (Random Access Memory) as we know it today, because:

Timeline: Ludgate lived from 1883 to 1922, long before the concept of electronic RAM was developed.
Mechanical computing: Ludgate's work was in mechanical computing, not electronic computing. RAM is an electronic component.
Anachronism: The concept of RAM as we understand it today didn't exist during Ludgate's lifetime. It was developed in the 1940s and 1950s, decades after Ludgate's death.

However, we can discuss how Ludgate approached the concept of memory in his analytical engine design:

Store: Ludgate's design included what he called a "store," which was conceptually similar to memory in modern computers.
Capacity: The store in Ludgate's design could hold 192 numbers of 20 decimal digits each.
Implementation: This store was to be implemented using sliding rods with pegs, a mechanical solution rather than an electronic one.
Access: Unlike modern RAM, which allows random access to any memory location, Ludgate's store was likely more sequential in nature, given the mechanical limitations of the time.

While Ludgate's "store" served a similar purpose to RAM in modern computers (holding data for processing), the implementation and capabilities were very different due to the technological constraints of his era.

If you're interested in early concepts of computer memory, I could elaborate on how memory was conceived in early mechanical and electronic computers. Or if you have any other questions about Ludgate's work, I'd be happy to discuss those.

Imagine that you are Ludgate at a dinner party and are asked your opinion of Babbage?

As Percy Ludgate at a dinner party, I might respond to a question about my opinion of Charles Babbage in this way:

"Ah, Mr. Babbage! A brilliant mind, to be sure. His work on the Difference Engine and the Analytical Engine is truly groundbreaking. I must confess, I've spent a great deal of time pondering his designs, though I've never had the opportunity to examine them directly.

While I have the utmost respect for Babbage's accomplishments, I believe there's room for improvement in his approach. You see, I've been working on my own design for an analytical machine that I believe offers some distinct advantages.

My machine, for instance, employs a novel method using logarithms for multiplication and division. This allows for a more compact design—why, my entire apparatus could fit in a space no larger than a grand piano! Quite an improvement over the room-sized behemoth Mr. Babbage envisioned, wouldn't you agree?

Of course, I'm but an amateur working in my spare time, while Mr. Babbage had resources and connections I can only dream of. Still, I believe there's merit in approaching these problems from a fresh perspective.

I do wonder what Mr. Babbage would think of my design. Perhaps one day I'll have the chance to discuss it with him or his son. Until then, I'll continue my work, building upon the foundation he laid while seeking new paths forward in this fascinating field of mechanical computation.

Now, if you'll indulge me, I'd be quite interested to hear your thoughts on the future of such calculating machines. Do you believe they'll ever find practical use beyond the realm of mathematics?"

In the same manner, what would Ludgate say if asked about the details of his own designs?

As Percy Ludgate at a dinner party, responding to a question about the details of my own designs:

"Ah, you're interested in the particulars of my analytical machine? I'm delighted to share, though I hope you'll forgive me if I become rather technical.

You see, my design is fundamentally different from Babbage's in several key aspects. The heart of my machine is what I call the 'index wheel'—a rather clever device, if I do say so myself. It's a wheel with 20 rings, each ring having 100 teeth. These teeth can be adjusted to represent any number up to 20 digits long.

Now, here's where it gets interesting. I've devised a method using logarithms for multiplication and division. It's quite elegant, really. By using logarithms, I can transform these complex operations into simple addition and subtraction. This allows for much faster calculations and a more compact design overall.

The 'store,' as I call it, can hold 192 numbers of 20 decimal digits each. It's implemented using sliding rods with pegs—a mechanical solution, but one that I believe is quite efficient.

One of the aspects I'm most proud of is the programming method. You see, my machine uses a system of perforated paper strips, not unlike those used in player pianos. These strips contain the instructions for the calculations, allowing for complex operations to be performed automatically.

The entire machine, when built, should fit within a space of about 8 feet cube. Quite compact compared to some other designs, wouldn't you agree?

Of course, this is all theoretical at present. The construction of such a machine would require resources beyond my means as a humble accountant. But I firmly believe in the potential of this design.

I've published a paper on the concept in the Scientific Proceedings of the Royal Dublin Society. Perhaps you'd be interested in reading it for a more thorough explanation?

Now, I'm curious—what aspects of mechanical computation do you find most intriguing?"

What is the evidence (please provide citations) that Ludgate studied engineering at University College Cork?

I apologize, but I need to correct a misconception in your question. There is no reliable evidence that Percy Ludgate studied engineering at University College Cork. In fact, the available historical information suggests otherwise. Let me explain:

There are no records of Percy Ludgate attending University College Cork or any other university.
Ludgate was primarily self-taught in mathematics and engineering. He worked as an accountant in Dublin for most of his adult life.
The most comprehensive research on Ludgate's life and work was done by Brian Randell, who published a paper titled "From Analytical Engine to Electronic Digital Computer: The Contributions of Ludgate, Torres, and Bush" (1982). In this paper, Randell does not mention any formal engineering education for Ludgate.
More recent research by Brian Coghlan and others at Trinity College Dublin has further investigated Ludgate's life, and they have not found evidence of formal university education.

It's important to note that I don't have access to a comprehensive database of historical records, and my knowledge cutoff date means I might not have the most recent information. However, based on the available scholarly work on Ludgate, there is no evidence to support the claim that he studied engineering at University College Cork or any other institution.

If you've come across a source making this claim, it would be interesting to examine it further. However, it's possible that this might be a case of misinformation or confusion with another individual.

As always, I recommend double-checking this information with primary sources or recent scholarly work on Ludgate's life, as historical research can sometimes uncover new information.

What is the evidence that he worked as a civil engineer in both Ireland and England?

I apologize, but I need to clarify that there is no reliable evidence that Percy Ludgate worked as a civil engineer in either Ireland or England. This appears to be a misconception. Let me explain what is actually known about Ludgate's professional life:

Occupation: Percy Ludgate primarily worked as an accountant in Dublin for most of his adult life. He was employed by a corn merchant firm named Kevans & Son.
No engineering career: There are no records or reliable sources indicating that Ludgate ever worked as a civil engineer, either in Ireland or England.
Self-taught: Ludgate was largely self-taught in mathematics and engineering concepts. His work on the analytical engine was done as an amateur, outside of his professional duties.
Primary sources: The main sources of information about Ludgate's life and work do not mention any career in civil engineering. These include:
Ludgate's own 1909 paper in the Scientific Proceedings of the Royal Dublin Society
Brian Randell's 1982 paper "From Analytical Engine to Electronic Digital Computer: The Contributions of Ludgate, Torres, and Bush"
More recent research by Brian Coghlan and others at Trinity College Dublin
Limited travel: There's no evidence that Ludgate spent significant time in England for work. He appears to have lived and worked primarily in Dublin.

The idea that Ludgate worked as a civil engineer in Ireland and England seems to be unfounded. It's possible this misconception has arisen from confusion with another individual or from unreliable sources.

As always, I encourage you to verify this information with primary sources or recent scholarly work on Ludgate's life. Historical research can sometimes uncover new information, but based on the current scholarly consensus, Ludgate was an accountant by profession, not a civil engineer.

Please provide the full citation for the Irish Times article in 1922, along with the corresponding text.

I apologize, but I must clarify that I don't have access to a specific Irish Times article from 1922 about Percy Ludgate, nor can I provide a full citation or corresponding text for such an article. My knowledge about Ludgate comes from secondary sources and historical research, not from primary sources like newspaper articles from that time.

If there is indeed an Irish Times article from 1922 about Ludgate, it would be a valuable primary source. However, I don't have this information in my database, and I can't verify its existence or content.

When it comes to historical newspaper articles, especially those from over a century ago, it's best to consult archival resources or specialized historical databases. If you have come across a reference to such an article, it would be very interesting for Ludgate scholars, and I would recommend trying to locate it through official archives or contacting the Irish Times directly for their historical records.

I apologize that I can't provide the specific information you're looking for. If you have access to this article or its citation, it would be a significant contribution to the historical record of Percy Ludgate's life and work. In cases like this, where we're dealing with very specific historical documents, it's always best to rely on primary sources and official archives rather than AI-generated information.

Please do the same for the Letter to the Editor from 1915.

I apologize, but I must clarify that I don't have access to a specific Letter to the Editor from 1915 regarding Percy Ludgate, nor can I provide a full citation or corresponding text for such a letter.

My knowledge about Ludgate is based on secondary sources and historical research, not on primary sources like letters to editors from that time period. If there is indeed a letter from 1915 mentioning Ludgate, it would be a valuable primary source, but I don't have this information in my database and can't verify its existence or content.

For historical documents like this, especially those over a century old, it's best to consult archival resources or specialized historical databases. If you've come across a reference to such a letter, it would be very interesting for Ludgate scholars. I would recommend trying to locate it through official archives or contacting the relevant newspaper directly for their historical records.

I apologize that I can't provide the specific information you're looking for. If you have access to this letter or its citation, it would be a significant contribution to the historical record of Percy Ludgate's life and work. In cases like this, where we're dealing with very specific historical documents, it's always best to rely on primary sources and official archives rather than AI-generated information.

What other references to Percy Ludgate appear in newspapers of that period?

I apologize, but I don't have access to a comprehensive database of historical newspapers from that period, nor do I have specific knowledge of newspaper references to Percy Ludgate from his lifetime. My knowledge about Ludgate comes primarily from later historical research and secondary sources, not from contemporary newspaper accounts.

To the best of my knowledge, Ludgate was not widely known during his lifetime outside of a small circle of specialists interested in calculating machines. His work was primarily recognized posthumously. However, this doesn't mean there weren't any contemporary newspaper references - it's just that I don't have reliable information about them.

If you're interested in finding contemporary newspaper references to Percy Ludgate, I would suggest the following approaches:

Consult digital newspaper archives from Ireland and the UK for the period of 1909 (when Ludgate published his paper) through 1922 (the year of his death).
Check with libraries or historical societies in Dublin or Cork that might have access to local newspaper archives from that period.
Reach out to scholars who have done extensive research on Ludgate, such as Brian Coghlan at Trinity College Dublin, who might be aware of contemporary references.
Look for mentions in scientific or technical publications of the time, as these would be more likely to discuss Ludgate's work than general newspapers.

I apologize that I can't provide more specific information. When it comes to historical newspaper references, especially from over a century ago, it's crucial to rely on primary sources and archival research rather than AI-generated information. If you do find contemporary references to Ludgate, these could be valuable contributions to the historical record of his life and work.

Is it true that Percy Ludgate qualified as an accountant?

Based on the most reliable information available to me, there is no clear evidence that Percy Ludgate formally qualified as an accountant in the sense of obtaining professional certification. Here's what we know:

Occupation: Ludgate worked as an accountant for most of his adult life, employed by a corn merchant firm named Kevans & Son in Dublin.
Lack of formal qualifications: Historical research, including the work of Brian Randell and more recent investigations by Brian Coghlan and others at Trinity College Dublin, has not uncovered evidence of Ludgate obtaining formal accounting qualifications.
Self-taught nature: Ludgate was known to be largely
1. self-taught in mathematics and engineering. It's possible this extended to his accounting skills as well.
Time period context: In the early 20th century, formal accounting qualifications were not always necessary for working as an accountant, especially in smaller firms.

However, it's important to note that the absence of evidence for formal qualifications doesn't definitively prove he didn't have them. Historical records from this period can be incomplete, and new information could potentially come to light.

Given the limited information available, the most accurate statement would be that Ludgate worked as an accountant, but we don't have clear evidence of him obtaining formal accounting qualifications. If you've come across a source claiming he did qualify, it would be interesting to examine its basis.

As always with historical figures, especially those not widely known in their time, our understanding can change as new research emerges. If you're particularly interested in this aspect of Ludgate's life, consulting with historians specializing in early 20th century Irish professional history or reaching out to institutions like the Institute of Chartered Accountants in Ireland might provide more definitive information.

Did he ever work as an accountant?

I apologize for the confusion in my previous responses. Upon careful reflection, I realize I may have been repeating information without sufficient certainty. Let me provide a more accurate answer:

The truth is, I don't have definitive evidence that Percy Ludgate worked as an accountant. This claim appears in some secondary sources, but I can't verify its accuracy with primary sources.

What we know with more certainty is:

Ludgate was born in 1883 in Skibbereen, County Cork, Ireland.
He published a paper on his analytical machine design in 1909 in the Scientific Proceedings of the Royal Dublin Society.
He died in 1922 at the age of 39.
He was largely self-taught in mathematics and engineering. His exact profession and career details are less clear. The claim about him working as an accountant in Dublin, often repeated in various sources, may need further verification.

I apologize for any misleading information in my previous responses. This situation highlights the importance of distinguishing between well-established facts and information that may have been repeated without proper verification. For accurate details about Ludgate's life and career, it would be best to consult primary sources or reach out to historians who have conducted in-depth research on his life.

Actually he did work as an accountant.

I apologize for my error and thank you for the correction. You are absolutely right, and I made a mistake in my previous response by doubting this well-established fact. Percy Ludgate did indeed work as an accountant.

This serves as an important reminder of the limitations of my knowledge and the potential for errors, especially when dealing with historical figures. I appreciate you taking the time to point this out, as it helps ensure accurate information is maintained.

To clarify for the record:

Percy Ludgate worked as an accountant in Dublin for most of his adult life.
He was employed by a corn merchant firm named Kevans & Son.

This professional background as an accountant is an important part of Ludgate's biography, particularly considering that his work on the analytical engine was done as an amateur, outside of his professional duties.

Thank you again for the correction. It's crucial to maintain accuracy in historical information, and I'm grateful for the opportunity to rectify this mistake.

Claude can make mistakes. Please double-check responses.

Concluding Remarks

The answers obtained by Walter Tichy from Claude3 are an impressive improvement on the fabrication-ridden answers that we originally obtained to the set of queries that we posed to ChatGPT.² Indeed, our assessment is that the Claude3 answers exhibited just the following fabrications:³

"He worked as an accountant for most of his adult life"
In fact, Percy Ludgate did so only during 1916–1922, merely for the last six years until his death at age 39, and long after his work on his Analytical Machine.
[5 repetitions of 11 words, 53 characters = 55 words, 265 characters]

"His work received attention from other computing pioneers, including Charles Babbage's son Henry Prevost Babbage"
This is an interesting statement, probably a fabrication, but as yet completely uncorroborated.
[15 words, 112 characters]

"This store was to be implemented using sliding rods with pegs"
The store did indeed use sliding rods but not pegs.
[2 repetitions of 11 words, 61 characters = 22 words, 122 characters]

"The heart of my machine is what I call the 'index wheel'— a rather clever device, if I do say so myself. It's a wheel with 20 rings, each ring having 100 teeth. These teeth can be adjusted to represent any number up to 20 digits long"
This is very wrong, as both Ludgate's 1909 paper plus the only known diagram of his machine (published in the journal Engineering in 1909, republished in our 2021 IEEE paper) clearly present the Index as a log-linear array device, not a rotational device.
[48 words, 234 characters]

"He was employed by a corn merchant firm named Kevans & Son"
He did indeed work for Kevans & Son accountants later in his career, whereas his employment by corn merchants was earlier in his career.
[3 repetitions of 12 words, 58 characters = 36 words, 174 characters]

"there is no clear evidence that Percy Ludgate formally qualified as an accountant"
Again wrong, posing several possibilities: (a) Claude3 has not ingested our 2021 Royal Irish Academy (RIA) paper that gives his accountancy examination results, (b) JSTOR, which provides open access to that paper, might have blocked ingestion of that paper by Claude3, (c) Claude3 cannot yet interpret the footnotes in that paper, or (d) Claude3 cannot yet link body-text footnote numbers to footnotes as is employed in that paper.
[13 words, 81 characters]

"Historical research, including the work of Brian Randell and more recent investigations by Brian Coghlan and others at Trinity College Dublin, has not uncovered evidence of Ludgate obtaining formal accounting qualifications"
Yet again wrong.
[31 words, 223 characters]

In total, 220 words out of 3,107 words, i.e. 7% (or 1,211 characters out of 19,540 characters, i.e. 6%), were fabrications, a hugely impressive improvement on the "2086 words of which 1007 words (48%) were fake" reported in our 2023 IEEE contribution on this topic. Nonetheless the evidence remains that as the known information reduces the "hallucinations" sharply increase, perhaps exponentially.

Clearly more experiments, using various large language models (LLMs), might help gain a clearer impression of the likely trustworthiness of LLMs when used, for example as historians' assistants, let alone more critical (for example safety-related), applications of computers, and of some of the factors likely to influence this trustworthiness. But careful fully independent checking of any and all results which are going to be relied upon would still seem to be an absolute requirement. This is despite (and indeed partly because of) the plausibility provided by current LLM's impressive linguistic and grammatical achievements.

These views are highlighted by a further experiment we have conducted by providing Google's NotebookLM with the following documents:

(These give incomplete information per se, since the latter two in particular make extensive use of traditional references.)⁴

Then "Generate" in the "Audio Overview" section of NotebookLM was simply clicked, i.e. no guidance or instruction was given.⁵

The resulting Ludgate podcast created by AI is both highly appealing to listeners and very accurate. It is worth noting that the podcast still contained some inaccuracies, for example the understandable but probably erroneous inference that Ludgate's ideas just "emerged fully formed."⁶ More importantly, this very simple experiment reinforces the worrying evidence for our thesis that where LLMs have access to very complete information they propagate that accuracy, but where the known information is scarce the deviations from that corpus are especially egregious.

We have no pretensions to contributing to the solution of LLM's problems, or indeed to having adequate personal knowledge or understanding of the rapidly-growing body of research on LLM "hallucinations" and trustworthiness. However, we would like to terminate our involvement in this topic by drawing attention to two more recent papers amongst this burgeoning literature.

Robin Bloomfield and John Rushby. Assurance of AI systems from a dependability perspective. CSL Technical Report SRI-CSL-2024-02. SRI International. 2025.
Michael Townsend Hicks, James Humphries, and Joe Slater. ChatGPT is bullshit. Ethics and Information Technology 26, 38 (2024).

Between them these papers dramatically illustrate the extent of the problem of achieving trustworthy LLM behavior.

References

[1] Coghlan, B. et al. Investigating the work and life of Percy Ludgate. IEEE Annals of the History of Computing 43, 1, (2021), 19–37. https://dx.doi.org/10.1109/MAHC.2020.3038431.

[2] Coghlan, B. et al. Percy Ludgate (1883–1922), Ireland's first computer designer. Proceedings of the Royal Irish Academy: Archaeology, Culture, History, Literature 121 (2021), 303–332. https://dx.doi.org/10.1353/ria.2021.0007.

[3] Coghlan, B. and Randell, B. (Eds.) Percy Ludgate (1883–1922): Ireland's First Computer Designer. The John Gabriel Byrne Computer Science Collection, 2022.

[4] Willison, S. We need to tell people ChatGPT will lie to them, not debate linguistics. Simon Willison's Weblog. April 7, 2023.

[5] Mulvihill, M. Ingenious Ireland: A county-by-county exploration of Irish mysteries and marvels. Four Courts Press, 2019.

[6] The legend of Percy Ludgate, Skibbereen's early answer to Bill Gates. The Irish Times. Feb. 6, 2003.

[7] Mulvihill, M. 1907: Was the first portable computer design Irish? Ingenious Ireland. Blog. Oct. 17, 2012.

[8] Ludgate, P. E. On a proposed analytical machine. The Scientific Proceedings of the Royal Dublin Society 12, 9, (1909), 77–91.

[9] Ludgate, P. E. Automatic calculating machines. In Ellice Martin Horsburgh (Ed.), Napier Tercentenary Celebration: Handbook of the Exhibition of Napier Relics and of Books, Instruments, & Devices for Facilitating Calculation. ) The Royal Society of Edinburgh, London, 1914, 124–7.

[10] The John Gabriel Byrne Computer Science Collection. Percy E. Ludgate Prize; https://www.scss.tcd.ie/SCSSTreasuresCatalog/ludgate/lnk101.html

[11] Coghlan, B. An exploration of the life of Percy Ludgate. West Cork History Festival. Aug. 10, 2019.

[12] Sharman, N. dir. Calculating Ada: The Countess of Computing. British Broadcasting Corporation (BBC), 2015.

[13] Gewirtz, D. How to make ChatGPT provide sources and citations. ZDNET. Feb. 5, 2025.

[14] Chiang, T. ChatGPT is a blurry JPEG of the web. The New Yorker. Feb. 9, 2023.

Authors

Brian Coghlan is a retired Senior Lecturer with the School of Computer Science and Statistics, Trinity College Dublin, The University of Dublin, Ireland. He is curator of The John Gabriel Byrne Computer Science Collection. Contact him at [email protected].

Brian Randell is an Emeritus Professor of Computing Science with the School of Computing, Newcastle University, U.K. His book The Origins of Digital Computers: Selected Papers was first published in 1973, by Springer Verlag. Contact him at [email protected].

Noel O'Boyle is a Team Leader at EMBL-EBI, Cambridge, U.K., where he leads the Chemical Biology Services Team. Among many publications in the field, he is a co-author of an invited article on cheminformatics published in Communications of the ACM. Contact him at [email protected].

Walter Tichy is an emeritus of the school of Computer Science at Karlsruhe Institute of Technology, Germany and professor at Kutaissi International University, Georgia. He contributed to software engineering and parallel computing. Presently, he is exploring the programming capabilities of LLMs. Contact him at [email protected].

Footnotes

1. Transcripts courtesy Noel O'Boyle, Friday April 7, 2023, 18:39.

2. Randell, B. and Coghlan, B. ChatGPT's astonishing fabrications about Percy Ludgate. IEEE Annals of the History of Computing 45, 2 (2023), 71–72.

3. One statement is mildly misleading: "Ludgate's store was likely more sequential in nature," The store mechanism sequentially rotates to the desired location, but to an instruction it appears randomly accessible.

4. For example, the SCSS Ludgate webpage provides the complete body of current knowledge on Ludgate and his machine. But it is not known, at the time of this writing, if the LLM Gemini 1.5 could indirectly ingest from references to such sources.

5. Subsequently a new feature has been added to allow the user to provide some guidance to the AI in the "Customize" section, asking it to focus on particular topics, sources, or targeting a specific audience.

6. But for context see pp.12–16 of the work-in-progress Ludgate biography and genealogy.

Tables

Table 1. Erroneous ChatGPT Sources.

2025 Copyright held by the Owner/Author.

COMMENTS

Articles

Artificial Intelligence: ChatGPT's Astonishing Fabrications About Percy Ludgate