Unveiling The Untold - Inside the Complexities and Controversies of ChatGPT

A headline written by ChatGPT

? Jeffrey Robinson 2023

These three things I know as true:

1) Yesterday, every man, woman and child with an email address was being targeted multiple times a month by professional fraudsters.

2) Yesterday, one in eight people was losing money each year as a result of those targeted frauds.

3) Today, artificial intelligence (AI) is making the criminal landscape catastrophically much worse and absolutely no one is safe.

The AI generated photos of Taylor Swift that appeared on X, formerly known as Twitter goes far beyond anything that should be tolerated. In certain jurisdictions it's a criminal offense. Luckily for her, she’s wealthy enough to have lawyers who will fight the windmills. Hopefully, further litigation will follow. But several things about this worry me. The ease in which this can happen. The anonymity of those who perpetrate such crimes. And the people who who shrug, why care about Taylor Swift? Those folks are completely missing the point. Will they be equally blind to the dangers posed when some angry boyfriend targets their teenage daughter?

Shockingly, deep fakes like that are fast becoming our new reality. In the New Hampshire primary for the Republican ticket in late January, deep fake robo calls mimicking President Joe Biden’s voice were used to discourage voters from coming to the polls. It’s easy to guess who paid for them, but no suspect has yet been named. Nor is it clear what crime has been committed. But guaranteed, short of legislation, deceiving voters this way will happen more and more as the election season heats up.

Understandably, as technology explodes into every aspect of our lives, any technology that enhances our lives is, perhaps, a good thing. I say perhaps, because you cannot honestly assess a technology’s true benefits until you weigh it against the side effects. Much like diet pills that advertise you can lose 40 pounds in a month, check out the small print – non-stop diarrhea for the first two weeks followed by two weeks of painful constipation, bladder leakage, liver damage, loss of hair and permanent damage to your hearing and eye sight. Same goes for the small print that should appear somewhere with new technology.

Case in point, one such “perhaps” AI end-product called ChatGPT.

It’s a program, currently used by more than 100 million people, that can write anything for you. Authors, like myself, worry that publishers will fall for it. Apparently, a computer generated novel cannot be copyrighted, so that's a good thing. But if an author puts his/her name to it? Now we're in the land of quicksand. Plagiarism is already a major problem, as is the theft of intellectual property, as is the stealing of what's called "the sweat of the author's brow." But ChatGPT doesn't obey those laws. Guns don't kill people, people kill people, but a world without ChatGPT is one where computers can't kill a writer's career. Hollywood screenwriters went on strike last year because they feared, rightly, that film and television producers will fall for the benefits of ChatGPT, realizing that they can get scripts written for free.

How good is ChatGPT? I asked it to compose one paragraph on the poem Humpty Dumpty in the style of an eight year old.

“Once upon a time, there was a funny egg named Humpty Dumpty. He sat on a wall, all high and tall, looking so jolly. Then one day, whoopsie-daisy, he fell down, and oh my goodness, he made such a loud sound! All the king's horses and all the king's men rushed to help him, but poor Humpty was broken, oh dear! But don't you worry, because in the end, they managed to fix him up again, and Humpty was happy as can be, sitting on that wall, smiling with glee. The end!”

?Besides the fact that all the king’s horses and all the king’s men couldn’t put Humpty together again – proof that ChatGPT is more than capable of outright lying – it just as effortlessly wrote one paragraph on the same poem when I asked it to mimic a PhD candidate.

“The narrative of Humpty Dumpty, often underestimated for its simplicity, harbors deeper allegorical implications worthy of academic exploration. Embedded within its seemingly whimsical verses lies a profound commentary on the fragility of human endeavors and the inevitable nature of downfall. Humpty Dumpty's precarious perch atop the wall symbolizes the precariousness of human ambition and the hubris that often accompanies it. His fall, although comically depicted, serves as a poignant reminder of the inevitable consequences of overreaching. Moreover, the subsequent futile attempts by the king's horses and men to restore Humpty Dumpty to his former state underscore the limitations of human agency in the face of irreparable damage. Thus, beneath its nursery rhyme fa?ade, Humpty Dumpty offers a compelling meditation on the complexities of human existence and the inexorable march of fate.”

On the surface, perhaps, some of it might seem innocent enough. We all knew a few people in high school and college who actually read the assigned books, but we also know plenty more who took the easy way out with Cliff Notes.

As a comparison, here’s the way Cliff Notes begins its report on Willa Cather’s novel My Antoinia:

?“The narrator meets a childhood friend, Jim Burden, now a successful lawyer for a railroad company, while on a train trip crossing Iowa, and they reminisce about growing up in the same small Nebraska town. When Jim says he wonders why the narrator has never written about ántonia, the narrator makes a pact with Jim that she will write about ántonia if he will. (In some editions of the novel, Jim is already writing about ántonia when he meets the narrator.)”

So why bother going down the well-trodden Cliff Notes path, when ChatGPT is fresh and new and can make you sound that much more intelligent?

To wit: “Willa Cather's novel My ántonia stands as a testament to the enduring power of the American frontier and the immigrant experience in shaping the nation's identity. Set against the vast and unforgiving landscapes of the Nebraska prairie, Cather intricately weaves together the lives of her characters, particularly the resilient immigrant girl ántonia Shimerda, to illuminate themes of love, loss, and the pursuit of the American Dream. Through vivid prose and poignant storytelling, Cather captures the essence of pioneer life and explores the complexities of cultural assimilation, friendship, and the bonds of community. As one of the most celebrated works of American literature, My ántonia continues to resonate with readers, offering profound insights into the human condition and the transformative power of the American landscape.”

Which one gets the better grade from the unsuspecting 11th grade English teacher? That’s obvious. Still, ask yourself, honestly, which one fails the student who chooses to cheat? The answer is, clearly, both of them. It’s just that ChatGPT makes cheating easier, faster and more readily available.

Now consider the world of scientific and medical research where lives can be at stake. How about scientific papers on some new cancer treatment being faked by ChatGPT to help secure tenure for a doctor when the result is AI bullshit?

Expanding the scope of such cheating, I would argue that ChatGPT and AI are the keys to a kingdom of fake documents that look absolutely real (no more cut and paste utility bill photocopies); false transaction and financial statements that look totally genuine; and invoices, contracts and deal memos that will fool just about everyone, especially bankers and business executives willing to hand you money based on those documents.

Here, for instance, is what ChatGPT came up with when I asked it to write letter from Amazon’s Jeff Bezos to JPMorgan Chase chairman and CEO Jamie Dimon introducing and recommending the famous bank robber Willie Sutton:

Dear Jamie,

I am delighted to introduce to you Mr. Willie Sutton, a distinguished individual whose remarkable talents and experiences have greatly impressed me. As the founder of Amazon, I have had the privilege of encountering many exceptional individuals, and I can confidently say that Mr. Sutton ranks among the most extraordinary. With a background spanning [mention any relevant details about Willie Sutton's background], Mr. Sutton brings a unique perspective and invaluable insights to any endeavor he undertakes.

Throughout his career, Mr. Sutton has demonstrated a remarkable ability to [highlight specific skills or achievements relevant to the recipient]. His dedication to [mention any relevant causes or industries] is truly inspiring, and I have no doubt that his contributions will greatly benefit your organization.

I wholeheartedly recommend Mr. Sutton and am confident that he will excel in any role he assumes. Please do not hesitate to reach out to me if you require any further information regarding Mr. Sutton or his qualifications.

Warm regards, Jeff Bezos Founder, Amazon

Starting to get the picture?

A recent study undertaken by the Dawes Centre for Future Crime at University College of London late last year identified a long list of future crimes. “Criminals can impersonate individuals through convincing audio or video manipulation, potentially leading to financial fraud or the manipulation of public opinion; the advent of autonomous vehicles presents an opportunity for terrorists to carry out coordinated attacks without human involvement; phishing attacks (which) can craft highly convincing messages, making it difficult to distinguish between genuine and malicious communications; (where) systems become integral to various sectors, criminals may target them, causing chaos, power failures, or financial disruptions; large-scale data harvesting and personal vulnerability identification, making traditional blackmail scalable; and, fake news (which) can manipulate public perception, though it may not always directly yield financial profit.”

The Center concluded, “While AI has the power to revolutionize various industries, its misuse poses significant challenges to society. Law enforcement, policymakers, and technology developers must stay vigilant and proactive in addressing these emerging threats to safeguard our increasingly AI-dependent world.”

A Europol study went even deeper into three specific crimes.

* Fraud and social engineering: ChatGPT’s ability to draft highly realistic text makes it a useful tool for phishing purposes. The ability of LLMs to re-produce language patterns can be used to impersonate the style of speech of specific individuals or groups. This capability can be abused at scale to mislead potential victims into placing their trust in the hands of criminal actors.

* Disinformation: ChatGPT excels at producing authentic sounding text at speed and scale. This makes the model ideal for propaganda and disinformation purposes, as it allows users to generate and spread messages reflecting a specific narrative with relatively little effort.

* Cybercrime: In addition to generating human-like language, ChatGPT is capable of producing code in a number of different programming languages. For a potential criminal with little technical knowledge, this is an invaluable resource to produce malicious code.?

Looking at the future is one thing. Thinking long term and big picture, then acting on that, is something many law enforcement officials, politicians, bankers, lawyers and others who have a stake in the future of financial crime, seldom do.

Perhaps ChatGPT looks almost harmless in the hands of a 15 year old who needs a book report tomorrow morning. It isn't, but he doesn't yet understand why. And yes, all social advancements and all technology bring with it its own dangers. Automobiles kill horses, pedestrians and children playing in the streets. Sitting too close to a television can cause young boys to become sterile. Computerized voting machines can be manipulated by Venezuela’s dead strongman Hugo Chavez.

But then there is this warning from ChatGPT itself:

“As adults engage with ChatGPT, it's crucial to remain mindful of its capabilities and limitations. While ChatGPT can provide valuable information and assistance across various topics, it's essential to remember that it's an AI language model and not a human being. Relying solely on ChatGPT for critical decision-making or personal advice may lead to misunderstandings or incorrect conclusions, as it lacks human intuition and empathy. Furthermore, being mindful of the potential biases in the data it was trained on can help prevent the perpetuation of misinformation or harmful stereotypes. Utilizing ChatGPT as a tool for information gathering and entertainment can be enriching, but exercising discernment and critical thinking is paramount to ensure a safe and beneficial interaction.”

I highlight those three words, “can be enriching,” to set up the small print caveat for the word, perhaps. And here I add ChatGPT’s own definition: “Perhaps is an adverb used to indicate possibility, likelihood, or uncertainty about something. It suggests that something is possible or might happen but is not certain. It's often used when expressing doubt, speculation, or a conditional statement."

*****

There’s plenty more where this came from. You can find a growing collection of pieces on Substack at https://substack.com/@writingfactory.

The site is free to browse and free to subscribe to.

I am also available, and would be delighted if you’d join me, on the brilliant Twitter-alternative BlueSky at https://bsky.app/profile/writingfactory.bsky.social

Finally, if you’ve enjoyed this, please follow along on Linked In.

Even better, send an invite to connect on Linked In and I will happily accept. I’d be flattered if you did and would welcome your company

?

?

?