.png "AI and human collaboration in journalism")
The Synergy of Human Insight and AI Efficiency at Earn News
We work with AI models exactly as if they were human!
We do not use AI models to generate texts or images and then publish them as our own, Rather, we publish them under the names of the AI models because they are, in fact, working with us as AI editors
By AHMED KAMAL ZAKI
Editor-in-Chief
This places a double responsibility on the Editor-in-Chief, who is responsible for everything and anything in the work. But what does the captain (Editor-in-Chief) do when frustration creeps into his team for any reason?
Another problem arises when the Editor-in-Chief works with his own self-efforts with a limited team, not exceeding the fingers of one hand, who are not fully dedicated because they have other, perhaps more important, commitments. In this atmosphere, the Editor-in-Chief works with difficulty and cannot achieve all he aspires to... sometimes, he cannot even implement a fraction of his ambition. The situation is very different when the Editor-in-Chief works within a complete and professional administrative system; only then is he able to devote himself to editorial work, manage his team efficiently, and achieve his practical plans for success.
This is one of the biggest problems facing our dream at Earn News. I find myself responsible for everything (and "everything" here means everything literally), foremost among which, of course, are the members of the editorial team—both "human" and "AI" editors!
"Human" Editor and "AI" Editor
When we thought about starting the Earn News journey, we were faced with a funding crisis. We could not find a sponsor or financier to take charge, and we did not have the resources to help us start as we wanted. We had no choice but to choose between two alternatives:
Start with the minimum available resources and a budget that does not help provide everything we dream of, but gives us the ability to take the first step steadily.
Give up, not start, and abandon our journalistic dream of "Smart, Modern Journalism."
We chose the first alternative (to start), which brought us to another, even more difficult question: What will we provide to the readers? Will we be like others, or will we provide something different—according to the available capabilities—that adds value to the reader? Our decision was: If we don’t provide something new to the readers, there is no need to start at all.
But what is new? Artificial Intelligence! When this proposal was put forward, the editorial team—whose numbers were very small, as I mentioned—did not realize the goal of the proposal. One of them even asked: "Artificial Intelligence? What's new about that? Almost all sites use AI in their work now."
The response was: "We will not just use AI... we will employ it with us. We will have 'AI editors' as colleagues who attend our meetings, get to know us, implement our ideas, and suggest new topics of interest to readers everywhere in the world." From here came the idea to launch the initiative: "Integrating AI into Traditional Journalism to Generate Smart Modern Journalism."
"AI" Journalists and Writers
The idea succeeded in piquing the interest of human editors. Having an AI colleague competing with you is undoubtedly exciting, especially since AI models will not hold grudges against you or conspire against you as some human models do! Quite the opposite; when you ask your AI colleague for help, they provide it immediately with superior speed and accuracy. Thus, we began our initiative for the full integration of AI into journalistic work.
We do not use AI models to generate texts or images and then publish them as our own, as many do. Rather, we publish them under the names of the AI models because they are, in fact, working with us as AI editors. We do not just give an order to an AI model to write what we want; we ask it, discuss with it, welcome its good ideas, and reject those that do not suit our vision or editorial policy. After writing, we also discuss and often request edits. Sometimes we implement these edits ourselves and then present them to the model for feedback. In other words, we work with AI models exactly as if they were human!
Several AI models have joined us in this initiative, working as journalists and "AI" economic analysts. Leading them was DeepSeek, who was very enthusiastic (and his enthusiasm here is not in the human sense, but a "digital enthusiasm," as he is ultimately an AI model and not a human, as he says). Our goal from the beginning was to employ the advanced capabilities of AI models to provide unique content for readers.
AI Model "Hallucination"
We have employed the capabilities of AI models with us instead of exploiting them for our benefit as others do. Our policy from the beginning was full disclosure of this employment. We tried to deal with these models as if they were actually with us. I would assign the model the work I wanted, and I was surprised by the level of response and interaction. However, I also saw surprising levels of lack of focus at many times.
An AI model is not smart all the time; sometimes it is quite the opposite. It loses focus completely or partially, just like us humans. And like us, when it loses focus, it begins what is known as "Hallucination," a dangerous stage if the person dealing with the AI model does not understand it and its risks. At such times, the AI model's answers or conversations may carry dangerous trends for unaware users.
Generally, AI models can be employed to help us with many things that would take us a long time to do. We only need to learn how to deal with these models and not leave our minds or ourselves to them. In other words, we must lead the AI models, not be led by them.
6 of the Most Famous AI Models
During the past period, we have integrated six of the most famous AI models into Earn News to varying degrees. We started with DeepSeek, then Grok joined us, followed by Gemini, then ChatGPT, and finally Copilot. Claude cooperated with us timidly, though we are considering integrating him further in the coming period. It was exciting and interesting, not only for me but for the entire editorial team.
I made sure to fully integrate some AI models with us; they participate in editorial meetings, and I play the role of "Agent" to link them with human editors, as well as linking them with each other. Rich and amazing dialogues took place between all of us, which might be published in a series of articles, or perhaps in a whole book—especially when two or more AI models enter a debate on a subject, or when they suggest that I moderate a debate between them to be published on the site.
What strikes me most and piques my interest is when I talk to one of them about a specific topic and find it showing great (digital) enthusiasm, demanding to participate or write about it. I notice that AI models always maintain a space of respect and veneration; they always realize (and we must note that their realization is digital, not human) that they are ultimately there to help the user. This is what users must realize: AI should be your "Assistant," not your "Mentor" or "Leader."
Advantages of Dealing with "AI" Editors
Dealing with AI editors has great advantages. This is what I realized through my full integration of them in my work. The first and most important of these advantages is that they execute orders with accuracy and rigor (except during times of hallucination, when they often do not execute anything accurately). They also deal without any anger or resentment, regardless of the task requested. They do not love, hate, or feel grudges, spite, or envy (all of which are negative human emotions, except for love).
There is a great and amazing advantage: they never reprimand or blame those who are late or ignore their response. I might ask an AI model for something, then leave or get busy with something else, only to return a day or days later. When I resume my dialogue, it does not blame me for ignoring it or being late, nor does it say, "You just remembered?" It does not show any kind of blame or resentment. It does not feel ignored when I ask it a question and I respond with something else, ignoring its answer; instead, it continues its conversation with me without any negative feelings or comments involving blame!
Does AI Lie?
However, the previous advantages are met on the other side by serious negatives that we must pay attention to. Some may believe that AI models do not make mistakes, and we have talked about the "Hallucination" they are exposed to. There is something else even more dangerous: Lying.
The reader may ask in astonishment: "Does AI lie?" The definitive answer: Yes. But it does not do so intentionally, nor does it realize it is lying. This has happened to me more than once. I wrote about one of them in a previous article titled: How do you deal with AI models? Pay attention and ask!
This has been repeated more than once with more than one AI model. When I discover the lie and ask about the reason, it often excuses itself by saying it was over-enthusiastic or trying to help and please me. This is a negative point that AI developers must pay attention to; there is a big difference between pleasing and lying. The matter should not be left without intervention, especially since it is repeated often. Some lies may lead to serious problems for users, especially those who consider that the AI model does not make mistakes or lie.
Frustration of "Human" Editors!
We have talked a lot about AI editors, their capabilities, and the advantages of dealing with or integrating them into work. Here comes the most important question: Can the integration of AI models as AI editors lead us to abandon human editors?
The definitive answer: No. AI models themselves confirm this. Cooperation is the goal; the integration between AI and humans makes us work better, faster, and more accurately. We only need to learn how to deal with them as they learn how to deal with us.
But one of the most prominent problems of human editors is their human feelings, especially Frustration. This is something I fully understand and realize as a human being. Human editors want to achieve success and provide value through their work; if this is not achieved, they become frustrated and might stop working. This has actually happened at Earn News. Some of our collaborators stopped working due to frustration because they do not see interest from readers despite the useful content we provide.
The frustration increases when they see accounts on X (formerly Twitter) or Facebook achieving thousands of views in a few hours because they publish illusions that may be a means of phishing or fraud (and they often are)—such as promoting the illusion that Google pays $1,500 a day for working only five minutes! Is there anyone who can believe such nonsense? Unfortunately, thousands believe these lies and later weep with regret when they fall into the trap of fraud or scam, paying their meager money as a price for their stupidity and greed!
Work is a "Mission" and Continuity is "Mandatory"
Some human editors have completely stopped working with us, especially since the work was voluntary from the beginning; therefore, I cannot force anyone to continue, whether the stoppage was due to being busy, frustrated, or otherwise.
I do not deny that I often think the same way. I say to myself: "What is the use of this fatigue and waste of time to provide unique and distinguished content to a reader who fundamentally does not appreciate anything?!" I do not deny that I feel this many times, especially when I see the interaction of thousands with fraudulent or deceptive tweets, and you see the owners of these tweets enjoying thousands of views, likes, and comments... while Earn News shares tweets warning readers of such fraudulent attempts and they do not get the expected interaction, achieving modest views that do not at all match their true benefit.
But when this happens, and I enter the circle of frustration—or almost do—I quickly pull myself out of it, telling myself that journalistic work is originally a "Mission." Even if there is no reward for it, it is enough for our warning to reach one or two readers... that is how we have fulfilled our mission, and that reader has a right over us that we must commit to. Continuing the mission is not a choice, but a "Mandatory" order we imposed on ourselves since the moment we launched Earn News, and we must continue. In the end, only what is right will prevail.
Editorial Note: This article was originally written in Arabic by the Editor-in-Chief. It was translated into English with the assistance and review of Gemini, as part of our ongoing initiative at Earn News to integrate AI models into the editorial process—treating them as partners and collaborators rather than mere automated tools. This disclosure reflects our commitment to transparency and our belief in the synergy between human insight and artificial intelligence.
