From RAG to Agentic RAG for Faithful Islamic Question Answering

Gagan Bhatia, Hamdy Mubarak, Mustafa Jarrar, George Mikros, Fadi Zaraket, Mahmoud Alhirthani, Mutaz Al-Khatib,, Logan Cochrane, Kareem Darwish, Rashid Yahiaoui, Firoj Alam · 2026

Read From RAG to Agentic RAG for Faithful Islamic Question Answering by Gagan Bhatia, Hamdy Mubarak, Mustafa Jarrar, George Mikros, Fadi Zaraket, Mahmoud Alhirthani, Mutaz Al-Khatib,, Logan Cochrane, Kareem Darwish, Rashid Yahiaoui, Firoj Alam (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-07528.txt.

Description

Gagan Bhatia and a large team study how language models answer Islamic questions. Errors here carry serious religious weight, so grounding matters more than fluency. Standard multiple-choice tests miss key real-world failures, the authors argue. Models may invent free-form answers or fail to abstain when evidence is lacking. Their 2026 preprint moves from plain retrieval to agentic retrieval.

Retrieval-augmented generation supplies source text to ground an answer. An agentic version lets the system plan searches and check evidence in steps. The team examines faithfulness, meaning answers that stay true to sources. They also value proper abstention when sources do not support an answer. The work speaks to builders of religious question answering tools.

Catalog record

CREATORGagan Bhatia, Hamdy Mubarak, Mustafa Jarrar, George Mikros, Fadi Zaraket, Mahmoud Alhirthani, Mutaz Al-Khatib,, Logan Cochrane, Kareem Darwish, Rashid Yahiaoui, Firoj Alam
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE66 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-07528

Browse the full library · About the Tower of Babel