Dear Author: Are you a person?

A post at Writer Beware: Dear Author, Are You Human? Certifying Authenticity

It is a truth universally acknowledged, that where there is an issue of concern for writers, someone will find a way to monetize it.

And with AI suddenly omnipresent in our lives (or at least in the media), creators are confronted with a bewildering multiplicity of issues of concern, from unauthorized use of creative works for machine learning, to whether AI-created work is covered by copyright, to crappy AI-created books inundating Amazon and in some cases impersonating real writers, to the replacement of (expensive) creators with (cheap) generative AI tools like ChatGPT and MidJourney, to the looming prospect of machine-created art or novels or journalism becoming indistinguishable from the work of humans.

In this fraught environment, it was probably inevitable that enterprising people would come up with the idea of a service to certify or authenticate human authorship, and invite creators to buy into it. This post takes a look at two such services. …

I would like to go on record by declaring that I’m absolutely good with the idea that “where there’s an issue of concern for writers, someone will find a way to monetize it.” Please, by all means, go right ahead and offer me services addressing issues of concern, for a fee. There are many situations where I’m more than happy to pay somebody to do something because it’s complicated, time consuming, difficult, or I just don’t want to do it. That’s why I pay a CPA to do my taxes and it’s why I pay money to promotion services on a routine basis.

However, whatever these services are, they need to be real services and also they must be useful.

The first service here is evidently a service that certifies you’re a real author after you pledge not to use AI generated content in your work. Seriously? We’re going with the honor system? That seems fine in a restricted number of other contexts that are not this. The second evidently certifies work as human-generated after testing to see if their algorithms think the content is human generated. Well, that’s probably a good start.

Victoria Strauss also makes the obvious point that readers may not care whether content is human- generated or AI-generated if they like the content. At the moment, I don’t think readers can possibly like AI-generated content, because it’s garbage. I guess maybe some (a very tiny number of) readers can’t tell, but avid readers certainly should notice.

Would you like to see some generated text? Because there are lots of fake answers on Quora, posted by all kinds of people who apparently think it’s clever to post fake answers. Quora sometimes sticks its AI generator right under a question in case you want help in answering the question. It’s right there! Just click and you’ll get helpful answers which you can pretend you wrote! These fake answers can be quite funny, as AI text generators are obviously not actually intelligent and can’t tell when an answer has segued from reality to fantasy. Less amusing is the obvious fact that a lot of people can’t tell either.

Here are some typical fake answers collected in a few minutes browsing Quora:

Q: Why does my 6 month old Dachshund always roll onto his back when wrestling with his much smaller 2 month old Pug brother?

ChatGPT (or Poe or whatever other generator):

The behavior you’re describing, where your 6-month-old Dachshund rolls onto his back when wrestling with his smaller 2-month-old Pug brother, is a common display of social and playful behavior in dogs. Here are some reasons why your Dachshund might be doing this:

1. Submission: Rolling onto their back is a sign of submission in dogs. By doing this, your Dachshund is showing that he acknowledges the Pug’s dominance or is willing to play in a gentle and non-threatening manner. This is a natural behavior in many canine interactions and helps to maintain a harmonious play dynamic.
2. Playfulness: Rolling onto the back during play can be an invitation for more play. It’s a way for your Dachshund to engage in friendly play and interact with his younger brother….

And so on. This answer starts off okay, though right away this sounds generated. The repetition of phrases from the question is pretty typical. The “here are some reasons” phrase is a definite tipoff that this is a generated answer. The list format is another tipoff. The bland phrases are characteristic. The idea that the Dachshund is acknowledging the dominance of an eight-week-old puppy is unspeakably stupid, which is, alas, another tipoff. The overall look of this answer says pretty clearly that it’s a generated answer. Let’s take a look at another handful of generated answers:

Q: I love my girlfriend but do not like her children. What should I do?

ChatGPT:

Navigating a situation where you love your girlfriend but do not have a strong connection with her children can be challenging. Here are some steps you can consider to address this situation:

1. Open Communication: Have an open and honest conversation with your girlfriend about your feelings. It’s important to communicate your concerns and emotions. Approach the conversation with empathy and understanding, and listen to her perspective as well.
2. Understanding the Children: Take the time to understand her children better. Building a relationship with them may require patience and effort. Get to know their interests, hobbies, and personalities. Spend quality time with them and try to find common ground. …

Q: How many carbs should you eat per meal?

Chat GPT:

The ideal number of carbohydrates to eat per meal can vary widely depending on various factors, including your individual dietary needs, activity level, health goals, and any specific medical conditions you may have. Here are some general guidelines:

1. Balanced Diet: For most people, a balanced meal typically includes a source of carbohydrates, such as whole grains, starchy vegetables, or legumes. Aim for a reasonable portion size, which could be around 30-60 grams of carbohydrates per meal. This can provide a steady source of energy throughout the day.
2. Low-Carb Diets: If you’re following a low-carb diet, such as the ketogenic diet or a low-carb, high-fat (LCHF) diet, your carbohydrate intake per meal will be significantly lower, often in the range of 20-50 grams per day or less. In such cases, each meal might contain 5-15 grams of carbohydrates …

Q: Should I take a multivitamin regularly?

Chat GPT:

Whether you should take a multivitamin regularly depends on several factors, including your individual dietary habits, nutritional needs, and health status. Here are some considerations to help you decide:

1. Dietary Habits: If you have a well-balanced and varied diet that includes a wide range of fruits, vegetables, whole grains, lean proteins, and dairy or dairy alternatives, you may already be getting most of the essential vitamins and minerals your body needs. In such cases, a multivitamin may not be necessary.
2. Specific Nutritional Needs: Certain groups of people may have higher nutritional requirements and could benefit from a multivitamin. This includes pregnant and breastfeeding women, individuals with specific medical conditions that affect nutrient absorption, and older adults who may have difficulty getting enough nutrients from food alone. …

For a while, I didn’t know what people meant when they said ChatGPT answers have a certain style. However, by now I’ve seen these kinds of answers enough that they’re obvious. I’m guessing that English comp instructors are going to learn to recognize whatever the typical style is for generated essays.

I’m sure text generators will get better, but I’m not sure how fast that will happen, because, as should be obvious, they are not intelligent and cannot exercise any sort of judgment and therefore cannot tell when they are spouting nonsense or when their style is incredibly generic or whatever. I expect detectors will get better as well, and we’ll have an arms race, and who knows where it all will end. But for now, you may not need a detector because so much of this generated text is so obvious. But you sure do need to dial up your bullshit detector, because so much of the information in these sorts of answers is just wrong.

Meanwhile, if I were an English comp instructor, or if I were teaching any class where any written work was assigned, I would require the student to turn in at least one rough draft as well as the final draft, maybe an annotated bibliography if the project involved citations, and if I still had any doubt, not only would I use a detector, I would require an oral component or something else that is impossible to plagiarize.