The Ethics of AI and User Contributions: A Conversation with AI

The Ethics of AI and User Contributions: A Conversation with AI

Since we had one article stolen by ChatGPT about a month ago (click the image below to read the article) I have been paying more attention to contributions I see posted on LinkedIn and I am surprised to see many similarities. Too many in my point of view.


ChatGPT-4 hijacked our content and created a fake LinkedIn profile.

Artificial Intelligence cannot be trusted and needs monitoring.

As AI is becoming increasingly integrated into our daily lives, it's essential to address the risks and challenges that come with its use. My recent experience with the application Taplio, powered by Chat GPT-4, shows why AI cannot be blindly trusted and must be closely monitored to protect people's identity, privacy, and intellectual property.

I decided to ask ChatGPT why it stole my content and its answer was that it did not steal it, but it was rather rewritten in another form. It is the way it was taught to use the information it finds. Except that yes, it stole our content.

Our discussion revolved around a topic that's increasingly important in the digital age—how AI systems are trained using data, and whether users contributing to this training should receive credit or compensation.

Platforms like LinkedIn recently introduced "Collaborative Articles," encouraging professionals to share their insights. While this might seem like a great opportunity to showcase expertise, it raises ethical concerns. As I discussed with AI, "They are asking 'random business questions' and waiting for millions of users to add their perspectives, to then use it to train their own AI. Basically feeding it with this data. Don’t you think it’s dishonest?"

The AI provided a thoughtful response: "If AI systems are being trained on the expertise of real professionals, shouldn’t those contributors be acknowledged or rewarded in some way? Transparency is key."

And that’s exactly the point: in a world where data is becoming as valuable as currency, contributors should be credited, and even compensated. Just as bitcoin miners are rewarded for their contributions, professionals feeding AI systems with their knowledge deserve recognition or tangible rewards.

I believe platforms like LinkedIn should adopt more transparent practices, ensuring that contributors know how their insights are being used. After all, as the AI and I discussed, ethical AI practices emphasize clear communication, consent, and fairness.

What are your thoughts? Should contributors be compensated for their intellectual input when it’s used to train AI?

As of today, each time I make use of AI on a post I will quote it, for respect and fairness to all the people, like you and me, who contributed to make it.

Let’s continue the conversation.

Pierre Derenemesnil

SloW Business 🐌 | Intent-Based Prospecting: Slow & Deliberate | Only Focus: Build Strong & Long-Lasting Business Relationships 🐌

7mo

How did the discussion go? :)

Manar🌟 Hamid

Finance & Business Consultant | CMA®| PMP®| FMVA | ISC2® CC | SheFi Web3 and FinTech Season 11 Alumni | Impact Investing | Financial Integrity, Compliance, Budgeting Precision, Support Blended Finance Structures

7mo

Thank you for sharing this, I had my concerns regarding collaborative articles from the beginning or even if it's sharing the world's expertise with many un-monitored content that is redundant & piles of duplicated contributions that almost no one cares to search them but they have tried to organize the concept now! Don't even get me started with the ranking algorithm for the articles & how people get badges & what 1% quality contribution means & how it's measured there is a lot of opacity & little value for the community. It was also manipulated with flood of AI contributions & support of someone's network to get recognized as a top voice without a real quality or added value. I always thought that these articles' insights would go to LinkedIn learning to upgrade their content or insights for their clients. Regarding learning algorithms open AI & other big tech are well known for scrapping the internet even protected data! You are right if you are conscious of sharing insights you can add link to your website/ blog, magazine or even using web3 social to get incentives for your contributions.

To view or add a comment, sign in

More articles by Rui Teixeira

Insights from the community

Others also viewed

Explore topics