🐈
» Forums » Freelancers » False results for AI generated content
Page options
a_ivanovska
Community Member

False results for AI generated content

Content writers, does anyone have a problem with AI detectors marking their work as AI-generated when it's in fact your own original work? This has become a big issue for me in the last 2-3 weeks. How do you prove to a client that the work is yours?

20 REPLIES 20
melroen
Community Member

If this has become a recurring issue, why don't you email yourself a copy of your work-in-progress file after each work session?

 

That way you will have a time-stamped record of your evolving drafts to prove that it's your original work on hand if any client claims it's AI-generated. 

Helpful ideas.

 

Unfortunately, some clients want to avoid legal and ownership issues, and believe the way to counter poor tech is with more poor tech. As Clark mentioned, there is no accurate AI, actually chatbot, detector. Clients do not understand they only need to look for quality and accuracy, as assessed by a native speaker from the area. I  suggest all clients with concerns use a quality plagiarism detector.

Exactly there is no Accurate AI detector yet on the market, However, Clients don't know this all they see is AI detection from whatever software they are using and your guilty LOL

 

celgins
Community Member

<start rant>

I would be more concerned about a client who uses AI detectors to test my work.


If a client properly vetted a skilled writer and chose to hire the writer, there is no reason to check the writer's work. Granted, the client should ensure the final product meets their requirements, but verifying a deliverable meets their needs is totally different from checking for deceit.


I would immediately drop any client who pushed my content through an AI detector. If a client isn’t confident they will receive quality material after reviewing my profile, and inspecting my educational background, and observing the quality of my examples, then don’t hire me.


You have 62 jobs dating back to 2017, and it looks like all of them are for writing services. Add a B.Ed. in English Language and Literature, and a 100% JSS covering those 62 writing jobs? Please! There is no reason for a client to check for AI-generated content from someone like you.


Maybe things like this irk me because my education and background are not in journalism, communications, literature, or writing. I write extensively for my clients these days and I have for years, and I don’t appreciate writers with your education/background/skill being “checked” because a client has a shiny new AI tool at his/her disposal.

<end rant>

 

Aleksandra--this isn't necessarily directed at you; I'm just venting, really. I agree with Melissa that timestamping drafts through email might help. It would be even better if the client has a online platform where you can upload your working drafts; that way, he/she can see your progress. Not a perfect solution by any means.

 

We all know that AI-detectors are not perfect; they're not even close. False positives happen way too often. Many AI-detectors would likely mark our messages in this thread as AI-generated!

melroen
Community Member

If you convert them to PDF the file is timestamped when last modified. I suggested email because you can add to the email thread daily for a project and have the files accessible in one place if needed.

 

Maybe not the most convenient solution. However, with the amount of edits and revisions a writer goes through to produce a final draft to submit to a client, you will have irrefutable proof of your original work.

 

Maybe someone else has a more practical solution? It appears this might become an issue writers will face going forward now that so many non-professional writers are using AI-generated content and passing it off as their own.

 

Since this is becoming more prevalent in the community forum lately, it makes me think this is a new ploy ill-intentioned clients use to avoid paying for work.

 

 

celgins
Community Member

Since this is becoming more prevalent in the community forum lately, it makes me think this is a new ploy ill-intentioned clients use to avoid paying for work.

Exactly. A ploy by a rogue client to squeeze high-quality, original material from a professional writer with no intentions of paying.

I hired writers before. I check everything and do Copyscape for plagiarism. I wouldn't take it personally as a writer, if that is your website and you are ranked very well in the search engines that site is your baby, you should not trust anyone at first until you worked with them for a while. The owner of the website will take the hit from Google, not the writer 🙂 my two cents worth. 

That's an understandable perspective from the client's side about intellectual property rights.

 

Particularly since the US Copyright Office won't copyright AI-generated work. If a party doesn't own the IP rights, they can't use, publish, or sell the content.

 

I'm uncertain how IP rights are dealt with in other countries worldwide.

yofazza
Community Member

I don't know, but I think doing the writings in Google Docs will keep everything archived without any effort from the writer's side. We can make the client able to trace every single revision, and even check the text before they were revised, if they want.

melroen
Community Member

Good idea, Radia, for many types of content writing.

 

Except, Google Docs isn't compatible with all file formats.

 

For example, Final Draft screenwriting software doesn't format correctly in Google Docs. But you can save a clean copy of a script to a  timestamped PDF file of your work-in-progress.

 

Looks like there are a couple of options depending on what kind of content you write and what formats you write in to disprove a client's allegation of AI-generated content:

 

-Save to timestamped PDF.

 

-Google Docs archived documents

 

-Copyscape or another reliable plagiarism tool.

 

But the takeaway is a freelance writer of original content must be proactive in this new environment of AI-generated accusations.

 

Whatever method works the best for each writer--document your work process each step of the way.

Yes, I have a bulletproof solution. Turn on screen recording while you work. You can then simply say that you have a recording, take screenshots, clippings, and in a particularly bad case, show everything.

Or say before job stsrted that you will record it. This is enough to discourage the client from even thinking about arguing if he planned to do it to receive free works.

That's a great idea to dissuade potential bad actors. Mykola!

 

Better to dodge a bullet from the start--if they disappear once they see they can't run their scam on a savvy freelancer--than to have it end in tears.

 

a_ivanovska
Community Member

Thank you everyone for sharing your ideas and opinions. I believe combining a timestamped PDF/version history in Google Docs and a plagiarism checker should solve the problem. I should also not allow myself to be bullied by clients who want to get premium content while paying reduced rates. I must admit that since ChatGPT and other AI tools emerged, I find it much harder to find new clients, which is probably why I put up with situations like this. Thanks again.

Thank you, Alesandra, for bringing up this disconcerting topic as a warning to all original content writers. Forewarned is forearmed. And we did come up collectively with some good ideas.

 

I checked out your profile. With your track record on Upwork and credentials, any client would be fortunate to work with a freelancer of your caliber! Don't let the bottom-feeders bully you--you're highly skilled and top-notch.

 

Wishing you much success!

Thank you for your kind words.

crart
Community Member

I see AI nonsense got out of hand sooner than I thought.

a_ivanovska
Community Member

Speaking of which, does anyone else have trouble finding clients? I'll be honest, I've been applying for content writing jobs for more than two months now and haven't managed to land a single client.

I was having a pretty difficult time for the last half of last year and the beginning of this year.  It's not picking up again but pay is down. Seems to coincide with the AI trend. But just a theory.

Since Feb after increseing connects everyone facing this issue....low pay, no view by client, not that much invite

lebach858
Community Member

It just happened to me. So weird. I don't know if the AI detector was really used or client is trying to pull something weird.  Been having more weird and shady interactions lately in the last year or two.

 

And why is there no way we can contact customer service? So ridiculous. 

Latest Articles
Top Upvoted Members