Newsletter and Grammar Coach

We humans are an odd lot. We treasure the majesty of nature — the wondrous flora, fauna and geologic formations of the natural world. Yet ever since the dawning of the industrial age, many of us have been intent on seizing every opportunity to circumvent nature by concocting synthetic products for expediency, for the sake of obtaining something faster, easier and cheaper.

That insatiable urge gave rise to artificial flowers, synthetic Christmas trees, artificial food coloring and flavors, fake turf, imitation snow, artificial fingernails, synthetic sweeteners, faux bricks and paving stones, even artificial robotic animals and humans. The latest and most disturbing movement is the headlong rush to embrace artificial intelligence (AI), and specifically generative AI. The difference between the two is significant. “Traditional AI is reactive — focused on processing and analyzing data to provide predictions or insights. In contrast, generative AI is proactive — capable of creating something new using learned data patterns,” as the Massachusetts Institute of Technology explained in an article on machine learning that was published October 22, 2024. A Forbes magazine article published on July 23, 2023, distilled that more succinctly: “Traditional AI excels at pattern recognition, while generative AI excels at pattern creation.”

The establishment of the OpenAI artificial intelligence research organization in 2015 stimulated the breathtakingly fast propagation of numerous AI platforms for all kinds of applications, in a wide range of fields. AI has been adapted for use in manufacturing, medicine, education, engineering, aeronautics, finance, military strategy — and, yes, writing and editing. It crossed the consumer threshold in November 2022, with open AI’s introduction of the wildly popular platform ChatGPT, which became the fastest-growing consumer software application ever. The “GPT” part of the name is an abbreviation for “generative pre-trained transformer,” which is the term for a language model that generates text based on the semantic relationships between words in sentences. Other widely disseminated AI tools with editing functions include Jasper AI, ProWritingAid, GrammarCheck, QuillBot, Linguix, Ginger, Rubriq, Writesonic, Rytr.me, WhiteSmoke, Wordvice AI Writing Assistant, Inkforall, Language Tool, Grammarly, Wordtune, CrawlQ, Hemingway Editor and CopyMonkey — perhaps the most telling name of all.

Although the technology is new, the term “artificial intelligence” is not. It was coined at an academic summer retreat at Dartmouth College in New Hampshire in 1956, to further contemplate a concept about which computer science pioneer Alan Turing mused in 1950: can machines think?

The key problem — and you know there is one — is that AI is not truly intelligent, in the proper sense of that term. The word “intelligence,” which is derived from the Latin term for “understand,” cannot truly understand something in the way that the human brain can. To understand something is to perceive its intended meaning, and moreover, to comprehend the significance of something and to be sympathetically aware of the nature or character of something, such as colors, music or emotions. No, AI cannot replicate that human capability. All that AI can do is to create databases of content that is fed into it, including patterns and sequences of words from previously created documents, and then regurgitate that information when commanded to do so. AI is fundamentally a highly sophisticated data storage and organization mechanism. In the realm of written communication, AI is a digital plagiarist.

At EditPros, we dismiss AI as subordinate to GI and HI — our terms for genuine intelligence in the form of human intelligence. One alarming caution about AI is its ability to learn and assimilate fallacies as well as facts, without the ability to distinguish between them.

The peer-reviewed scholarly journal Science Editor, published by the Council of Science Editors, contained an article in its August 19, 2024, edition scrutinizing the capabilities of AI for manuscript writing and editing. The article, by Rachel Baron, reported on the findings of a panel who subjected AI text writing and editing capabilities to precise testing. Their scrutiny documented pervasive problems with generative AI.

“The main problem with its use for academic research is its tendency to ‘hallucinate,’ whereby it presents facts and even references that look plausible but are completely invented. Another issue with generative AI is that it tends to reproduce inherent biases and stereotypes that exist in the training dataset," the study found.

The problems with respect to editing were equally disturbing. The article reported:

• After processing more than a few hundred words, ChatGPT seems unable to cope and starts deleting large chunks of text and replacing them with single-sentence summaries.
• Tools usually offer various prompts to vary the level of editing, although we have not yet found one that offers the right balance between under- and over-editing.
• A great disadvantage of these tools is that the editing is done at the sentence level, which creates problems with inconsistency and repetition. For example, abbreviations are redefined each time they appear because the AI tool does not recognize they have already been defined. Concepts and definitions of terms are also re-explained and redefined, leading to a great deal of redundancy.
• These suggested changes can, of course, be rejected, but because they are enmeshed with the more useful changes, it takes a lot of effort to distinguish necessary from unnecessary changes and diminishes any time savings that might have been gained from AI editing.
• Human editors should aim to edit text in such a way that it retains the author’s voice and preserves the intended meaning. AI takes neither of these things into consideration. Prompts that ask the AI to edit lightly tend to do little more than correct gross errors, but allowing it free reign by asking it to, for example, “edit the following text,” inevitably results in a complete rewrite that leaves little trace of the author’s original style.
• Given a string of words that do not immediately express a clear meaning, an AI editor will make a best guess based on the probability of particular words following one another, rather than on some inherent sense or understanding of the text. Sometimes this will result in a perfectly written sentence that clearly expresses the author’s intended meaning. At other times, it will result in superficially correct nonsense, or at least a distortion of the intended meaning. Unfortunately, AI tools aim to please by always offering a suggestion, which does not necessarily equate to an improvement.
• A rather alarming tendency of AI editors is to take quotations out of quotation marks, lightly paraphrase them, and delete the citation. It hardly needs pointing out that this could put authors at risk of unintentional plagiarism.

The study panel also disclosed that AI applications pervasively introduced errors in subject-verb agreement in complex sentences, verb tense dependent upon context, and inappropriate deletion of linking words and phrases needed for smooth transitions.

While the data management and organization capabilities of AI are astounding, it lacks human critical thinking capabilities. Also of concern are privacy and security considerations. The confidentiality of AI editing tools remains questionable.

Elsevier, the world’s largest publisher of scientific, technical and medical research journals and books, urges caution in use of AI for text editing. The Elsevier website declares, “Though AI-based editors can help authors rewrite sentences and catch grammatical errors, it doesn’t mean that complete reliance on AI for academic writing is advisable. Human review remains essential.”

Echoing concerns about bias, Elsevier advises, "AI learns based on big data, mostly generated by humans. However, all of its training data is not screened and therefore, may have biased or inappropriate expressions in content involving gender, race, culture, etc. AI-based editors may inherently mirror these biases. In order to avoid such mistakes, AI-edited work still needs to be carefully reviewed and screened, manually, to avoid misrepresentation in cultural and other aspects."

Elsevier declared, “Editing is a fine art. It requires experience and understanding of a subject area and the nuances of academic writing. Professional language editing services can also help you effectively avoid the risks that AI modifications may bring.”

Refinements undoubtedly will be made to AI manuscript editing functions. In that regard, the Science Editor article concluded, “whatever improvements are made to AI editing models, we believe that intervention by a human editor will continue to be an essential step in maintaining the high-quality service that academic editors offer their clients.”

We at EditPros unequivocally concur with that assessment of the limitations of AI editing, and we believe it applies to any kind of writing — marketing, consumer information, and website content as well as scholarly publications. Use AI, if you will, as a supplement, not a substitution, for genuine human intelligence..

Bart Lasiter owes his life to retired mechanical designer Jim Guigli. Jim didn’t save Bart’s life. He brought him to life.

Bart Lasiter is a fictional character whom Jim created out of his longtime enjoyment of reading mystery novels. Following retirement from the Lawrence Berkeley National Laboratory in the San Francisco Bay Area, Jim began to take more active involvement in his love of mystery fiction by writing. In the process, he created the character Bart Lasiter for an entry he wrote for the Bulwer-Lytton Fiction Contest — in which he won the grand prize in 2006.

Driven by a calling to flesh out Bart’s character more fully, Jim began researching the environs of a fictional detective office in a real place — Old Sacramento, only a few miles from his suburban home. Known as the West End in the late 1800s, that waterfront area had begun a slow decline in the early 20th century, and slid deeper into blight during and after the Great Depression. It remained a slum until the late 1960s, when a preservationist movement initiated a slow restoration of the dilapidated buildings. As renovation progressed through the ’70s, it had become a swarming tourist destination by day and by night transformed into a carousing bar scene.

One of the shabbier buildings served as the seedbed for Bart’s haunts — his cramped, drab apartment and office digs to which he began attracting a roster of seamy clientele. Jim Guigli plunged himself into studying Old Sacramento to give as much dimension as he could to his narrative as he began writing about Bart Lasiter’s exploits, initially in short-story form.

But when he began work on a book-length narrative, he realized he needed assistance to transform his manuscript from a word-processing document to a published book. His search for help led him to EditPros’ BookPrep service. Our designer, Marti Childs, consulted with him through the manuscript development process, guided him in shaping the content, and designed the template, cover and other graphic elements of the eventual 420-page book, titled Under the Black Flag — Piracy Is Not a Victimless Crime.

Once the book was properly formatted as a digital file, it became available in both paperback and hardcover editions through Amazon, Barnes & Noble and other retailers. It soon won an award in a book contest, the judges of which declared, “This was a superb noir mystery, with all the necessary elements, great characters, excellent dialogue, and great descriptions of settings.”

Although Jim chose to publish Under the Black Flag only in print form, EditPros also formats and converts manuscript files to conform to e-book specifications for sale through Amazon (Kindle), Apple (iBooks), Barnes & Noble (Nook), and Kobo, by means of submission to IngramSpark’s e-book service.

This is the 64th book that BookPrep has helped authors produce during the past eight years. With BookPrep, authors retain all rights to their books, and collect 100% of sales royalties.

We invite you to LEARN MORE about the EditPros BookPrep service.

Barry S. wrote:

“I see lots of folks on TV (and off), who should know better, use the words ‘in regards to… with the ‘s,’ when they mean ‘in regard to…’ as an alternative to ‘in reference to…’ or “regarding….’ Although ‘give my regards to Broadway’ works, as the plural form relates to a salutation, it should not serve as a substitute for the proper use of the singular, ‘in regard to.’ Am I wrong? Is this turning into another of those deals where they bastardize the correct usage so much, the wrong way becomes acceptable?”

The grammar coach replies:

* * * * *

Are you perplexed by some aspect of grammar or word usage? Don't be shy! Ask the "grammar coach" at EditPros and we'll try to help — at no charge, just for the sport of it.

Do you know of a friend or colleague whose office can benefit from EditPros’ services? If you do, you may be able to earn a “referral reward” of up to $500 from EditPros.   

We invite you to submit your questions to the "grammar coach," and we welcome you to subscribe to this monthly newsletter — which we'll send to your e-mail address at no charge. We respond to all "grammar coach" questions personally, but delays may occur because we must place top priority on assignments from our clients. We appreciate your patience and your interest.

You also can change your e-mail subscription address. For an address change, please indicate your existing AND your new e-mail address. Thank you.

OUR PRIVACY POLICY for e-mail newsletter subscribers:

We do not use any commercial e-mail lists or automated mass-mailing programs, and we do not allow access to the list by anyone else for any reason. Our subscriber list is maintained by hand, and it is not for sale. We are protective of confidentiality because many of our readers are also clients of ours.