by Mohamed Ousman (Islamic Movement, Crescent International Vol. 54, No. 12, Sha'ban, 1446)
Recently social media has been abuzz with discussion and debate on the award winning tafsir, The Ascendant Qur’an (TAQ).
During the discussion, some participants resorted to the use of Artificial Intelligence (AI) in order to form an opinion about TAQ based on comparative analysis with other translations. Using AI to assess TAQ was a first time development during a social media debate.
This article will bring out the salient features related to the use of AI in analyzing TAQ during that discussion and will conclude with the risks associated with AI.
The first attempt by AI to compare TAQ with other works focused on the translation of Surah Al-Fatihah, verse 5. The standard literal translation reads: “You alone we worship and You alone we ask for help.”
AI erroneously stated that TAQ translated it as: “You alone do we commit ourselves to serve, and from You alone do we seek assistance.” This brought the data-set and the integrity of the AI process into question.
TAQ across its three streams: the tafsir, the advanced English translation of the meanings of the Qur’an and the English translation of the meanings of the Qur’an uses the phrase “To You alone do we conform, and from You alone do we ask for help”.
Based on its unfounded erroneous wording, AI concluded that “TAQ translations constitute, not a translation providing an English equivalent of the Arabic words, but a tafsir reflecting Imam Asi’s interpretation and explaining the deeper meanings and drawing connections to contemporary issues based on a specific interpretive framework”.
Subsequently contradicting itself, AI went on to assert, “a standalone translation would typically be a single volume, or perhaps a few volumes at most. The extensive nature of the 14-volume set clearly indicates that it’s a detailed commentary and interpretation of the Qur’an, with the translation serving as a foundation for that commentary.”
AI was unable to differentiate TAQ tafsir from TAQ translations. In his tafsir, Imam al-Asi elaborates on “ibadah” (often translated as worship) encompassing a much broader concept than ritualistic acts but also actively striving to establish justice in the world.
In a comparison of Yusuf Ali’s work with TAQ, while AI acknowledged “the difficulty to completely eliminate all traces of interpretation it concluded that, compared to Imam Asi, Yusuf Ali’s translation is significantly more literal.”
When AI analysed Surah Al-Baqarah, verse 3, according to Yusuf Ali, “Those who believe in the Unseen, and establish regular prayer, and spend (in charity) out of the sustenance We have given them”; AI stated, “believe” is a common and generally accepted translation of يُؤْمِنُونَ (yu’minuna); “establish prayer is a relatively literal rendering of يُقِيمُونَ الصَّلَاةَ (yuqimuna al-salat).”
AI supposedly contrasted this with Imam Asi, and stated, “Ali does not use phrases like “inner certainty” or add bracketed words that fundamentally change the meaning of the verse, as Asi does.”
Again, upon revisiting the actual translation (and the tafsir), TAQ did not use “inner certainty”, highlighting yet another erroneous quotation by AI which once again brought the data-set and the integrity of the AI process into question.
Secondly, AI asserted that “The Ascendant Qur’an: Realigning Man to the Divine Power Culture” expanded to a 14-volume set confirming that it’s a very comprehensive and in-depth tafsir.”
Again, the opposite is true. TAQ began as a tafsir launched in 2008. It subsequently expanded to include an advanced English translation of the meanings of the Qur’an launched in 2021 and the English translation of the meanings of the Qur’an launched in 2024.
Thirdly, at the third stage of deep linguistic and interpretive analysis, AI concluded, “TAQ is more than a simple translation.”
Strangely, it concedes that the basic ingredients associated with any granular translation as “conveying meaning beyond literal equivalence, stylish choices, and the underlying interpretive framework.”
Fourthly, in what AI described as a comprehensive “Level 6 Analysis: Deconstructing Meaning-Making in The Ascendant Qur’an,” it stated that “this level of analysis delves into the philosophical and cognitive processes inherent in translation, recognizing that perfect equivalence between languages is impossible.” It acknowledges the translator’s subjectivity as an inescapable factor and understands interpretation as constitutive of the translational act itself. Applying this to The Ascendant Qur’an, we see why it is definitively a tafsir.”
Again, at Level 6 Analysis, AI confused the translation with the tafsir. While AI described the intricacies associated with translation such as “the impossibility of perfect equivalence, a translator’s subjectivity, interpretation as constitutive of translation, the underlying interpretive framework all resulting in the creation of a new text.”
Interestingly, at Level 6 analysis, AI states, “Al-Asi’s work is driven, inter alia, by liberation theology.” Liberation theology emerged out of the separation of state and religion in the context of Christendom. Imam Asi does not recognize such a split in Islam, hence the ideology of Liberation Theology cannot exist in the Islamic paradigm. This phraseology does, though, highlight the subtle lens used to analyze data as leaning towards secularism.
Fifthly, Zvezda Server (Звезда Сервер) churned out an opinion that “TAQ tafsir by Imam Asi is largely accurate in its analysis.” While it went on to use a loaded expression that “there are a few minor errors and areas that could be clarified” it explained this as being in the realm of “English words is not always a straightforward substitution.”
This highlighted another gap in the AI analysis, viz. the use of an incomplete data-set. Had this AI package factored in the Publisher’s Foreword, Translator’s Foreword, Translator’s Preface, Mufassir’s Foreword and Editor’s Foreword across the various streams of TAQ, that clarification would have been noted.
Among its suggested improvements, AI suggested that clarification of technical terms like, “tafsir”, “exegesis,” and specific Arabic terms could enhance understanding and accessibility.
This showed the short-comings in the AI process. Had AI employed a complete data-set it would have noted the Glossary contained in TAQ tafsir and advanced English translation of the meanings of the Qur’an which explains Arabic terminology. This is supplemented by a number of recordings. TAQ English translation of the meanings of the Qur’an contains no Arabic text or Arabic transliterations.
Zvezda Server (Звезда Сервер), Meta AI and 2.0 Flash Experimental all concluded that “TAQ constituted a tafsir not a translation.”
Incidentally, Meta AI stated “TAQ’s translator, Muhammad Asad, employs a unique approach…”; while 2.0 Flash Experimental stated “TAQ by Imam Jamil al-Asi is …” This mis-naming constitute errors at the most elementary level.
Sixth, another AI analysis concluded that Muhammad Asad’s The Message of the Qur'an is a widely respected high-quality translation. In its analysis of Muhammad Asad’s approach, it stated “Asad’s approach goes beyond strict literalism. He aims to convey the conceptual meaning of the Arabic text in clear and contemporary English. This often involves paraphrasing or using more descriptive phrasing to capture the essence of the original.”
This highlighted an AI inconsistency. Asad’s translation and TAQ translation by Imam Asi adopts a similar approach yet the former is classified as a translation whereas TAQ is classified as a tafsir. Incidentally, TAQ tafsir uses Asad’s translation with minor adjustments.
In this regard, AI stated, “if we were to classify Asad’s work based solely on the translation, it would be more accurate to call it an interpretive translation rather than a purely literal one. While it’s not as overtly interpretive as Asi’s work, it still goes beyond a simple rendering of the Arabic text.”
In the various analysis, all AI tools failed to detect that in TAQ tafsir, Imam Asi used the translation of Asad.
Seventh, in comparing Sahih International to TAQ and Asad, 2.0 Flash Experimental stated “Sahih International makes very few interpretive additions within the translation. It generally avoids adding bracketed words or phrases that significantly alter the meaning. It is generally considered to be one of the more literal and accurate English translations of the Qur’an.”
This clearly demonstrated an inherent bias which would make for a valuable investigation. Outside the virtual reality, in real life Sahih International is considered a poor translation of the Qur’an into English not only within academic circles but also among average readers. It contains major distortions and contradictions. Had it not been for the support of petro-dollars, it may not have reached notoriety.
To its credit, a point made consistently by the different AI tools regarding Imam Asi was that his work is known for its focus on specific themes, particularly social justice, liberation from oppression, a critique of imperialism and zionism and the establishment of a just world order. He often connects verses to contemporary events and issues, applying the Qur’anic message to modern challenges.
Lastly, AI tools are not transparent about the data they used to form opinions. The various users of AI had been unable to clarify the non-availability of the data upon which their AI tools expressed opinions. Hence, this article will be subject to review once the clarification about the data-set is received.
If anything, the data upon which TAQ was analyzed appears to be biased, incomplete and/or inaccurate. As TAQ is yet to be launched digitally, the electronic data upon which AI analysis is dependent is lacking.
Notwithstanding, the first part of this article concludes with one of the better accounts rendered by AI with human adjustment.
(Part 2 looks at inherent weaknesses in Artificial Intelligence (AI) application because of its reliance on past texts. Thus, it is incapable of appreciating any original thought).