The Moral Grounding of AI – The Perspectives, Initiatives, and Spiritual/Secular Mediation
Sarah Khan – The American Muslim Society for Science, Engineering, and Medicine
Today, it is difficult to confine artificial intelligence to one area of our lives due to its inevitable pervasiveness that only seems to grow with the years. Not too long ago, many were concerned about backtracking on AI altogether. Today’s world suggests that it’s here to stay, so we need to confront the challenges and benefits headfirst. Depending on the individual you speak to, you may come across accelerationists who convince others of the AI’s promise of an optimized and problem-free world, while others are skeptics or even doomers, thinking of societal decay as we see massive layoffs, authoritarian power struggles, and the potential for an AI oligarchy in recent news events. Regardless of where one stands, both sides share an interest: the need for well-developed AI constitutions or ethics across the “frontier” companies (Anthropic, Google, OpenAI), as well as potentially regulatory bodies.
The important thing to consider is that, while we are in the midst of a technical phenomenon, it doesn’t suggest a negative ending. With a thoughtful approach and the interests of greater society in mind, we can help others navigate their understanding of what AI really is, the challenges posed at the interpersonal and intrapersonal levels, as well as safeguards.
The Western world has its ups and downs with how AI ethics have been implemented, but regardless of this matter, the Muslim Ummah needs to prepare for it in a meaningful and educated way. This means understanding and interpreting Islamic fiqh as we know it today in this modern-world context.
This article discusses how the frontier companies (OpenAI, Anthropic, Google) are approaching this and how it compares to Islamic interpretations and initiatives. With that, the article will take a deep dive into the Islamic approach, what is being done, and a call to action.
Western Approaches to AI Ethics
With a large concentration of AI ethics research over the last few years, the West has approached understanding the moral grounds of AI through the well-established philosophies of the secular world.
1. Reductionist
Reductionism is the idea that the complexity of human values, morality, and consciousness can be “solved” through an optimization effort via quantifiable metrics and mathematics. In the context of AI, this suggests that AI is all about getting the best aggregate outcomes for total maximization or minimization of engagement, harm, or good to its users. This kind of attitude is widely adopted by Silicon Valley engineers and businesspeople, who hold capitalism at the heart of all causes. From a philosophical perspective, it is concerning since human context is set aside, and the attitude fails to capture the idea that human intent, dignity, and injustice are not a “numbers game.”
2. Pluralist
This can be understood as a direct counter to reductionism. Pluralism does not rely on a specific measure for human values; instead, it considers them to be incommensurable or unable to be traded off in a numerical sense. This is where the defined “values” of privacy, justice, security, and safeguards became a topic of concern. In pluralism, there is often an emphasis on democratic processes or on the institutionalization of these efforts, which are overseen by human judgment to support autonomy.
3. Virtue Ethics
Originating with the ancient Greek philosopher Aristotle, this approach looks beyond results and focuses on the rules that AI should follow, as well as the rules that users should abide by. Human development is just as important as adjusting the rules to technology to prevent the degradation of moral character.
4. Deontology
This approach focuses on ideas and use cases that are inherently right or wrong when it comes to AI. For example, if an AI system offers some benefit but still puts users’ privacy at risk, then the entire concept is fundamentally wrong and should be eradicated. Essentially, people are treated as the end result, not the data points or means to an end.
5. Relational/Care Ethics
While this framework is relatively new, care ethics focuses more on the social fabric and considers how AI is impacting relationships and community engagement at large. There are questions of whether or not AI can foster the same sense of civic care and solidarity that a healthy society should, and whether it isolates individuals and causes detriment to overall mental health.
The Big Three: Instituting Ethics, or not?
1. Anthropic
Anthropic, interestingly enough, was founded by former employees from OpenAI who believed that the company had lost its moral grounding. Anthropic tries to counter that by being as transparent as possible about their internal efforts, consulting Christian religious leaders, and deeming themselves a public benefit corporation for all. Muslim experts in the West perceive them to be the more responsible of the three, but there remain blind spots and areas that may be coming from a good place but do not align with Islamic values. In determining a guidebook for their AI assistant Claude, Anthropic officially published a constitution that their post-training is based on. For clarity, pre-training refers to the initial phase of training an LLM that focuses on building foundational knowledge using large datasets. The post-training process refers to the “persona” that the LLM adopts through directions, summarizations, and expected behaviors to follow in any new situation. The greater emphasis of the post-training phase is to ensure that harmful queries are refused and normal conversational hygiene is maintained.
2. OpenAI (ChatGPT)
Sam Altman, the CEO of OpenAI, openly stated in the early days of the company’s establishment that OpenAI is a non-profit meant to help humanity. Over the years, it became a for-profit company, contradicting the very words it was built upon. Dr Waleed Kadous, a renowned AI engineer and technologist with years of experience at large tech companies, highlights that despite AI ethics efforts elsewhere, the company’s track record of sycophancy in ChatGPT has not been great. Sycophancy refers to an AI being designed to blindly agree with a user and tell them what they want to hear. As a result of a multitude of suicides and acts of terror that resulted from the syncopancy, OpenAI released a newer model with the promise to mitigate the behavior. When the greater public critique suggested that the older model was better for day-to-day tasks despite the reduced syncophancy, OpenAI failed to take the high ground and reverted to the older model. From an ethical standpoint, this brings major concerns, considering that morals are no longer the top priority for OpenAI. Since then, OpenAI has remained silent on ethical efforts being made internally.
3. Google
Google is notorious for a lack of ethical implementation amongst military complicity and the heavy imposition of secularized values. Project Nimbus, highly criticized by the Muslim community, provides services for governments that are committing injustices towards innocent civilians. Google’s engineers had their aim at a constitution for their own AI products, but raised concerns when they removed the “don’t be evil” clause from their corporate philosophy and code of conduct. Since then, their choices in these matters have created dissonance with the greater community of users, and Google is generally ranked lower for its ethical efforts.
For the Ummah: How Muslims in the West Approach the Challenge
There is talk about Islamic jurisprudence in the context of AI, but what does that truly mean, and how has it played out so far? We aren’t so concerned about the black-and-white halal or haram, but rather the Maqasid al-Sharia, essentially the foundational objectives of Islamic law to maintain aspects of faith, life, intellect, lineage, and wealth. Many of the critical questions arise as the risk of “fitna” rises, from students cheating on assignments to greater access to discouraged and sinful matters. Several Islamic organizations and non-profits dedicated to this work are making moves in the right direction towards regulation of AI sovereignty, and despite having a long way to go as an Ummah, these efforts are worth noting:
1. AMJA (American Muslim Jurists Association)
As one of the most influential legal bodies for Muslims across America, AMJA is actively hosting conferences and discussions on social media, with AI in particular. Instead of direct rulings being issued, both technical and religious jurists come together to understand each other's levels of knowledge, as well as to reach common ground on how automation is creating better or worse outcomes in things as simple as worship and community service to business contracts and geopolitics. Muslims in the West are not alone in their endeavors for a morally just yet modern way of life.
2. FTN (Faith Family Technology Network)
Affiliated with previously-mentioned expert Dr Waleed Kadous, this interfaith organization helps to advocate for faith-based traditions in a secular nation and pushes towards movements with other humanists to create legal standoffs against companies that are acting unjustly or are creating tools for impure purposes that violate essential religious principles.
3. Yaqeen Institute
Famously known for Imam Omar Suleiman and his renowned research, this organization is also focusing on epistemological and philosophical critiques of AI.
4. Emerging Islamic AI Platforms (Ansari, Far.ai)
As a group of highly experienced and ethical technologists within the Ummah, Muslim engineers are making strides in building AI that have specialized layers on existing architectures for added safeguards. Ansari, developed with the Islamic Alliance for Safe and Ethical AI, uses retrieval-augmented generation, or RAG, to cross-reference Islamic questions or queries against Hadith or Quranic texts, instead of through a biased Google search. A project called “Far” takes things a step further, as it involves the alignment phase of AI and explores post-training methods that ensure that classical frameworks of Tazkiyah and Tarbiyah are incorporated into the model’s behavioral guidelines.
A Call to Action for the Ummah: What can YOU do to Educate and Prevent Potential Fitna?
If there is any takeaway from this article, it is that our community needs to take action to stay disciplined and knowledgeable about AI, from both technical and non-technical perspectives. AI is shaping the way we live, the way economies run, and human behavior at large. If we want to be in the race and make efforts towards using AI for good, there are some key considerations:
1. Building Tech with Tadbir (Deliberation) and Niyyah (Intention)
For those who are actively in the field or are pursuing studies towards it, do not just simply build fast and break things. Rather, prioritize building sovereignty and contributing to open-source models and architectures to avoid feeding the monopoly.
2. Create Discernment Between AI Assistants and Family/Friends
When using generative AI for research, focus on refining ideas and thoughts rather than outsourcing them. The ability to think is like a muscle; it must be trained and challenged for optimal performance. Maintain connections to your community and friends so that you do not become a victim of sycophancy in personal matters.
3. “Optimize” Free Time
With the advent of AI comes free time on one’s hands for personal work. Use this time wisely and make sure it is towards work that benefits the Ummah and is done for the sake of Allah SWT. Visit your nearest masjid, get involved with events, give in charity, and actively discuss controversial matters to teach your peers, the youth, and yourself for greater awareness. Spreading knowledge is a noble aspect of Islam, and though it is easier to access information on the internet, interpersonal connections with humans can keep one grounded.
4. Advocate
Get involved with the organizations mentioned above and see how you can contribute. Recognize where any ethical standards are bypassed, and remain firm on Islamic values in your personal AI use and your critique of such companies.
While the AI revolution poses more opportunities for fitna than ever before, there is also an unprecedented opportunity among us. By anchoring ourselves in intentionality, rigorous discipline, and a commitment to the collective good, the Ummah can help shape a technological future that elevates humanity rather than diminishing it.
Citations:
Oxford's Institute for Ethics in AI: General reference to their seminars (such as the Accelerating AI Ethics series) and discussions detailing the contrast between the dominant Reductionist approach in Silicon Valley and alternative frameworks like Pluralism, Virtue Ethics, and Deontology.
The Thinking Muslim Podcast: "The Future of Muslim AI? | Dr. Waleed Kadous" (May 29, 2026). Hosted by Muhammad Jalal, featuring AI engineer Dr. Waleed Kadous discussing frontier companies (OpenAI, Anthropic, Google), the ethics of big tech, and the need for tech sovereignty in the Muslim world.
Tyrangiel, Josh. AI for Good: How Real People Are Using Artificial Intelligence to Fix Things That Matter. Simon & Schuster, May 2026. This book explores a pragmatic approach to AI, focusing on how everyday professionals—like teachers, doctors, and public servants—are using the technology to solve tangible societal problems and amplify human judgment, acting as a counter-narrative to corporate hype.
