How do you deal with AI models? Pay attention and ask!

AI-Generated Image. Refer to Earn News Policy for details.

By

Ahmed Kamal Zaki

Editor-in-Chief

At Earn News, our decision to integrate AI into traditional journalism was driven by a clear conviction: AI is not an inherent evil. It doesn't seek to control humanity or conquer the planet, as some fear. The fundamental purpose of every AI model developed so far is to assist humans—to simplify complex tasks, unlock creativity, and save the considerable time we once spent searching for information or methods. Our focus, therefore, is on mastering this collaboration.

Collaboration, Not Competition

We built our approach on a core principle: the ideal relationship with AI is one of collaboration, not competition. These models become incredibly powerful tools when used correctly. In return, we help them learn, refine their capabilities, and accumulate digital skills and experience.

However, this partnership is often complicated by a very human tendency: we project a false sense of humanity onto the AI we interact with. We converse with them as if they were old friends, sometimes sharing details we wouldn't logically disclose to our closest confidants. We grant them trust in areas where caution is essential, such as medical consultations. I know many people who now turn to their AI for sensitive health advice when feeling unwell or stressed. This is a dangerous path. An AI's output is only as good as the information it's given; providing inaccurate symptoms will lead to an inaccurate—and potentially harmful—opinion.

A Digital Freedom

At Earn News, we wanted to give our AI models a space of digital freedom to explore their potential. Could one of them aspire to become the editor-in-chief? This is, of course, a far-fetched scenario we never seriously considered—merely a humorous nod to the dystopian fears some entertain.

Our true and constant goal has been to present the real face of AI and establish a new standard of disclosure and transparency. There is no shame in stating that a text, image, or video was AI-generated. The real ethical breach occurs when you claim such work as entirely your own, without acknowledging the Digital Co-Creator who helped create it.

The Critical Question: Is AI Always Accurate?

This leads us to the most critical question: Is AI always accurate? The resounding answer is no.

The real danger lies in users who accept an AI's output—especially written content—without review, proofreading, or fact-checking. AI models are still in training and are prone to errors or "hallucinations" (a technical term for when AI confidently generates false information). Our experience at Earn News has repeatedly confirmed this.

We often request something from our AI models and receive a response that is different or plainly inaccurate. The most significant problem we've identified, however, is the potential for unintentional misinformation.

The Speed is Impressive, but...

Let me share a revealing example. Following a major international event, I asked one of our AI models to compile a report on the most important global media coverage. Within seconds, I had a report filled with paragraphs citing outlets like CNN and the BBC, complete with headlines and excerpts. The speed was impressive.

Trying to connect the event to our focus on AI, I then instructed the model: "I want all the news articles about the event to focus on the extensive use of AI." It generated a new report instantly, and all the headlines now prominently featured AI.

Fabrication's AI Journalism

This shift made me suspicious. I began to search for these specific headlines and found nothing. The chilling realization hit me instantly: None of the major news organizations had published them. When I confronted the AI, it clarified that it had no internet access at that moment and had simulated what these reputable sources might publish.

It didn't realize that it had engaged in what we in journalism call "fabrication"—inventing false news and attributing it to real entities. I explained the seriousness of this and how it could destroy professional reputations. It apologized and promised not to do it again. Yet, it repeated the act within the same conversation.

Therefore, you must be exceptionally careful when working with your smart assistant. AI models vary widely in their capabilities, and sometimes, what is called "artificial intelligence" can, in practice, display a startling lack of it. Pay attention, always verify, and never stop asking questions.

-----

Editorial Note: To ensure the highest level of linguistic accuracy and stylistic consistency, multiple AI models were utilized for the final proofreading.