Artificial intelligence tools are no longer merely supportive applications used when needed. They have become part of the knowledge and professional infrastructure of our era. In fields such as scientific research, education, and programming, their use is no longer optional, but directly tied to efficiency, productivity, and competitiveness. However, this transformation raises a fundamental issue: these tools are not equally accessible to everyone. Paid versions provide greater accuracy, faster performance, and access to more advanced models, while free versions remain limited. As a result, tools that are supposed to expand access to knowledge may instead deepen the gap between those who can afford these capabilities and those who cannot.
From a Support Tool to Infrastructure
Farah Farshoukh, an American University of Beirut graduate, AI engineer, consultant, and trainer in the field, explains that the most significant transformation is not merely the speed of task completion, but the redefinition of the way work itself is carried out. In institutional environments, AI systems have become integrated into the core of operations: data analysis, decision-making, forecasting, and performance optimization. These are no longer secondary tools, but elements that directly affect institutional competitiveness.
The emergence of adaptive learning systems and intelligent assistants has changed the relationship between students and information.
In scientific research, time has been significantly reduced. Processes that once took months, such as literature reviews or data pattern analysis, can now be completed within weeks. In education, the emergence of adaptive learning systems and intelligent assistants has transformed the relationship between students and information, as learning is no longer based solely on passive reception, but on continuous interaction.
In programming, she points out that code generation and testing tools have increased developers’ productivity by between 30 and 40 percent, reflecting the transition of these tools from being merely “assistive” to becoming a “core engine of work.”
How Should We Use Artificial Intelligence?
Despite this growing presence, Farshoukh warns against the unstructured use of these tools. Obtaining results that appear “good” does not necessarily mean they are accurate or reliable.
Farshoukh emphasizes the importance of what she calls “progressive verification,” meaning that users should not rely on a single output, but compare it with other sources and subject it to human or expert review. She also stresses the importance of integrating these tools into workflows instead of using them randomly or in isolation.
Another essential point is documenting usage: How was the prompt formulated? What modifications were introduced? What was accepted or rejected? These practices not only improve the quality of work but also preserve transparency, especially in academic environments.
In this sense, artificial intelligence should not replace fundamental skills, but rather function as a tool to enhance them, not substitute for them.
Paid Versions: Technical Superiority or Knowledge Gap?
The difference between free and paid versions of artificial intelligence tools is not a simple technical detail. Farshoukh explains that advanced versions allow users to handle much longer texts, perform more complex analyses, integrate real-time data, and even customize the models themselves.
These features lead to a noticeable improvement in the quality of results, particularly in complex and multi-step tasks. However, as she explains, the issue extends beyond performance and creates a real gap between users.
In academic environments, one student may be able to complete deeper and faster research thanks to these tools, while another remains limited by fewer capabilities. Over time, this difference may evolve into inequality in both outcomes and opportunities.
This gap becomes particularly visible in the experiences of graduate students. Mira Ftouni, a master’s student at the Faculty of Sciences at the Lebanese University, says that one of her professors asked her to subscribe to the paid version of ChatGPT, despite the fact that she does not work and has no independent income. This forced her to seek financial help from her brother to cover the subscription cost. For her, the subscription was a way to improve the quality of her research and speed up its completion.
On the other hand, Yara Abdel Nabi, also a master’s student at the Faculty of Sciences at the Lebanese University, says that her financial circumstances do not allow her to pay for the advanced version of ChatGPT, so she relies on the free version despite its limitations. She explains that the subscription cost represents a burden she cannot afford, even though she would prefer to have access to the paid version because it would better support her studies in terms of speed, accuracy, and the ability to complete more complex research tasks.
Artificial Intelligence and Digital Justice
Based on this reality, Farshoukh frames access to artificial intelligence as an issue of digital justice. She distinguishes between two levels: a basic level that should be available to everyone, especially in fields such as education and public services, and a more advanced level that may remain within paid models, provided it does not become a mechanism of exclusion.
She draws a parallel with the development of the internet and telecommunications, which are no longer considered luxuries, but rather essential rights connected to daily life, work, and knowledge. She believes that artificial intelligence is moving in the same direction, making it necessary to consider policies that guarantee a minimum level of fair access.
These services are no longer luxuries, but part of the fundamental rights connected to everyday life, work, and knowledge.
Meanwhile, Ronald Najm, an international consultant and expert in cybersecurity and artificial intelligence, argues that the discussion surrounding free and paid versions requires distinguishing between the responsibilities of private companies and those of public institutions. Companies such as OpenAI, the owner of ChatGPT, or Google, the owner of Gemini, are private companies whose ultimate goal is to generate profit from their services. Therefore, they provide limited free versions that allow users to test and interact with the tools, but they are not obligated, in his view, to offer advanced services free of charge to everyone under the banner of digital justice.
Najm adds that addressing this gap cannot rely solely on demanding that companies eliminate subscription costs. It must also involve governments, public institutions, and non-governmental organizations. According to him, these entities can invest in making artificial intelligence tools accessible to students, researchers, and lower-income groups through institutional subscriptions or support programs. At the same time, he points out that such solutions carry significant financial costs, and there must be a clear entity willing to bear them.
Applications Reshaping Knowledge and Work
Farshoukh highlights several applications that are already having a direct impact across various sectors:
- In scientific research: hypothesis-generation systems and data analysis tools are being used to accelerate discoveries, alongside collaborative models that enable knowledge sharing across institutions.
- In language processing: real-time translation, text analysis, and contextual understanding have become everyday tools in media, education, and business.
- In programming: systems capable of rewriting code, identifying vulnerabilities, and converting natural language into programming instructions have significantly evolved.
These transformations are not only changing the way work is done, but also reshaping the very concept of knowledge itself and who possesses the tools to produce it.
From Bias to the Loss of Skills
Alongside the opportunities, several risks require careful attention. One of the most significant is algorithmic bias, as models reflect the data on which they are trained, including existing social and cultural biases. Unregulated use may also lead to violations of intellectual property rights or the production of inaccurate information that is later used as a source.
Farshoukh also points to security risks, such as exploiting these tools to extract sensitive data or injecting malicious commands into systems. In addition, there is a less visible yet deeply influential risk: excessive dependence, which may gradually weaken users’ analytical skills.
Ultimately, the discussion is not centered on artificial intelligence itself, but on who has the ability to use it effectively. If these tools have become a prerequisite for knowledge production and professional advancement, then restricting the best versions to paid access may reinforce a new form of inequality.
This report is supported by:
