Image source: Getty Images
Over the past months, social media sites have been engulfed in heated debate about the role of artificial intelligence in answering religious questions, following the launch of the “Qaf” application in the first week of last August.
According to the definition posted on the app’s page in the App Store, “Qaf” is “an intelligent Islamic assistant designed for Muslims who want an easy way to ask questions, learn and reflect.”
One of the commentators on the application page on the
On the other hand, another commentator on social media expressed the opposite opinion, writing: “If you are a Muslim, download this app.”
In the midst of this controversy, the promotional video for the application, published by one of its founders, Abdel Latif Abdel Fattah, received millions of views, and sparked widespread interaction and mixed comments, ranging from welcome, warning, and criticism.
This controversy reflects a broader debate that begins with the question of the difference between the human mind and the “mind” of a machine equipped with huge amounts of information, and passes through the global competition for technical leadership, all the way to the status of Islamic knowledge in the digital age.
An inevitable reality
A few days after the launch of Qaf, Abdel Fattah published the following clarification:
“There seems to be a big misunderstanding about QAF being a tool for issuing fatwas. To be clear, our position is that artificial intelligence cannot issue fatwas.”
Most read end
An in-depth explanation of the most prominent events and topics, to help you understand the most important variables around you and their impact on your life
Episodes
End podcast
He also published an example of a user who asked a poll question on the application, and Qaf responded by explaining that it is merely a tool for searching and accessing information.
The developers of “Qaf” define the application as a “platform” that uses artificial intelligence and “helps Muslims learn, explore and understand Islam.”
The developers do not claim to have designed or trained their own language model. QAF appears to be more of a tool that relies on existing artificial intelligence models, while providing it with Islamic sources from its database, in an attempt to limit the influence of non-Islamic sources in the answers it provides.
Abdel Fattah had indicated in the introductory video published on the X platform that the answers provided by “Qaf” were “generated by artificial intelligence, and artificial intelligence may make mistakes.”
These assertions may reflect the awareness of those in charge of the application of the sensitivity of dealing with religious issues, and the extent of scrutiny and criticism that may face any platform that addresses issues related to beliefs and legal rulings.
On the other hand, some welcomed and praised the idea, based on the fact that Muslims already use general artificial intelligence tools that were not designed according to Islamic references or controls. They believe that developing applications based on reliable Islamic sources represents a step in the right direction, even if the experience is not complete.
Supporters of this trend say that linguistic models have become part of the daily reality for many people, and that the question is no longer whether they should be used, but rather how they can be developed and controls can be put in place that ensure taking advantage of their capabilities in expanding the scope of knowledge and providing access to it more easily and directly.
Between enthusiasm and limits
If we look at the opinions of specialists in jurisprudence, we find an approach that combines enthusiasm for benefiting from new technologies and participating in their development, and reservation regarding the limits of their use and risks.
Dr. Salem Abdel Jalil, the Al-Azhari scholar and former Undersecretary of the Egyptian Ministry of Endowments, said in an interview with BBC News Arabic that applications that aid research, including artificial intelligence applications, “served the people of knowledge a very great service,” and that they are “useful for the student of knowledge who possesses knowledge and a conscious religious background,” and that they are “one of the things that God has bestowed upon us, the specialists.”
However, it distinguishes between two types of use, and distinguishes between two categories of users. In his opinion, the user should have a “religious background and legal knowledge that enables him to distinguish between truth and falsehood.”
He adds: "The specialized scholar can benefit from it to the greatest extent, because he is able to evaluate what he reads and hears through it. As for the non-specialist, he can learn from it as a culture, and not as a legal ruling, fatwa, or adjudication of a dispute."
It seems that the increasing trend towards the digital space to ask religious questions is now preoccupying ancient Islamic institutions.
Last year, Cairo hosted a conference entitled “The Making of the Righteous Mufti in the Age of Artificial Intelligence,” and its opening interventions focused on what participants considered two urgent necessities: benefiting from artificial intelligence, and establishing the necessary controls for its use.
Dr. Muhammad Al-Duwaini said, in a speech he delivered before the conference on behalf of the Sheikh of Al-Azhar, Dr. Ahmed Al-Tayeb, that benefiting from artificial intelligence “is a legal necessity, because it is neither rational nor wise for a person to live absent the blessings of his time.” But he stressed that the machine cannot be independent in issuing a fatwa, because making a fatwa is an “industrious process” that requires “the ability to sense and sense human nature and its conditions.”
The other concern that emerged during the conference was the quest to move the Islamic world from “the position of the object to the position of the actor and influencer,” so that “our minds are not managed from beyond the seas and oceans,” according to Al-Duwaini’s expression.
Image source: EGYPT'S DAR AL IFTA
An enormous heritage...but not digital
But the task is not easy, as the Arab and Islamic world faces great challenges in the field of developing artificial intelligence.
Dr. Samer Rashwani, professor and researcher at the Center for Studies of Islamic Legislation and Ethics at Hamad Bin Khalifa University, is pessimistic about the reality of current Arab models of artificial intelligence.
Rashwani told BBC News Arabic that some countries in the Arab and Islamic worlds realized early the importance of this field, and have launched ambitious projects since the emergence of generative artificial intelligence techniques.
But he believes that the path of development has been hindered by many factors, including the “artificial intelligence war between the United States and China,” which led to restrictions on the export of advanced technologies, especially chips used to train and operate artificial intelligence models.
Rashwani, who is also working on a project to develop an artificial intelligence model that will serve researchers in Islamic studies, explains that the effects of these policies still exist, “despite repeated American announcements about amending them.”
He says that the result is that the currently available Arabic models are “weak, very weak,” and that most existing projects do not focus on building models from scratch or training them, as much as they rely on employing ready-made models and developing tools on top of them.
This means that there is still reliance, to a large extent, on prevailing business models. Although these models have witnessed remarkable development and have become more capable of “covering general Islamic knowledge topics,” as he put it, they still suffer from shortcomings that limit their effectiveness in serving specialized Islamic studies.
Rashwani explains this by saying: “The main challenge is that the percentage of Arabic materials on the Internet used to train major business models does not exceed half a percent. The other challenge is that the Arabic sources available on the Internet are not comprehensive and do not represent all eras of Arab and Islamic history, nor all its branches of knowledge. There is a great disparity in the representation of these branches on the Internet.”
According to Rashwani, there is still a “tremendous manuscript heritage” that has not yet been digitized, in addition to a modern heritage that extends over the past two centuries. He adds: "All of this needs to be digitized if we want to reach an initial stage of an Arabic linguistic model capable of a more accurate understanding of the Arabic language, Arab heritage, and Arab culture in general."
There appears to be a general agreement among many specialists in the Arab and Islamic worlds on the necessity of moving from the stage of fascination with artificial intelligence to the stage of actual participation in its development, in parallel with developing frameworks that regulate its uses and determine its limits.
Discussions about generative systems quickly lead to issues of inherent bias in the data on which models are trained, and bias that the algorithms themselves may reflect.
While some researchers in the Islamic context are preoccupied with the influence of Western hegemony on the prevailing models, another no less important challenge may emerge, which is the wide diversity within Islamic schools and curricula themselves, which may open a new door to controversy with the spread of specialized Arab and Islamic applications and models.
Find out more
Short clips
End of short clips
