An Islamic Perspective on Ethical Considerations for AI-Powered Exploitation of Space, Time and Religious Ritual

The following is adapted by our Founding Member Yaqub Chaudhary from his presentation in an AI&F-hosted workshop on integrating faith beliefs with ethics and business practices related to AI in the workplace at the virtual Faith@Work ERG Conference on February 11 in Washington, DC.

How to integrate faith values and work ethics is an important question for every person of faith. In Islam, there is a prophetic statement that says that God loves to see that whenever anyone undertakes a task, that they perfect it:

 

إن الله يحب إذا عمل أحدكم عملاً أن يتقنه

 

Hence, individuals working in organisations or under contracts should aspire to the highest standards of professionalism to produce the best outcomes in their jobs since this is religiously praiseworthy and counts as an act of worship.

However, one difficulty today is that digital technologies are applied in society at scale and with immediate effect and furthermore, there is increasing competition to be the first to bring new products and solutions to market, or to find new ways of creating revenue, even if that means testing the boundaries of what is legal. Consider, for example, the numerous shocking corporate scandals that have since receded from public attention such as the VW emissions scandal.

What is needed is for religious practitioners to bring the ethical principles they derive from their religious traditions to create a culture that is more than merely protecting corporate interests, public image or bluffing regulatory authorities.

Despite the clamour of ethics, the reality is that the corporate ethics discussion around AI often amounts to a mere spectacle – an ethics theatre with little to no substance, and even a distraction while companies continue to build short-sighted, harmful or outright malevolent systems. Even at best, how corporations  approach ethics is constrained by commercial and economic prerogatives.

The ethical concepts that have been emphasised in AI ethics tend to be those that are most easily operationalised and solved with technical solutions.  Such solutions ease the path to deployment without raising scrutiny.  They are unlikely to bring in wider considerations about care, mercy, balance, probity and other virtuous characteristics with respect to the well-being and flourishing of individuals and the natural world (Hagendorff 2020, 103).

The absence of an overarching moral framework means that in the process of developing and deploying new technologies, everyone involved is faced with the challenge of knowing what is definitively good from what appears to be good but will lead to major harm for individuals, society and the environment. The key problem for a person of faith is that one’s duties in the workplace may come into conflict with broader religious visions and teachings for societal flourishing.

For Muslims, there is an overarching moral framework in the form of revelation and the four schools of thought in Islamic law, known as the madhabs, which provide guidance for correct action in every possible circumstance.  Actions fall under one of five categories: obligatory, prohibited, recommended, reprehensible or permitted (i.e. neutral).

The default category is that everything is neutral or permitted in the absence of any specific verse or statement indicating a command or prohibition. Hence, what matters is one’s intention – that one does not set out to do what is prohibited, and that what matters in every action is its quality in terms of consciousness of God and sincerity.

However, with modern technology, even where intentions may seem to be good all round, the complexity of new technologies means they unleash chains of unintended consequences or may be quickly repurposed for malevolent uses and applications such as:

Military uses, automated propaganda, propaganda and disinformation campaigns, social control, surveillance, facial recognition and sentiment analysis, social discrimination and exclusion, job losses and displacement, test deployments without informed consent, data breaches, unfairness and bias, deployments of harmful and flawed systems, undermining autonomy by manipulating users and microtargeting, the creation of echo chambers, filter bubbles and so on.

What is needed is deeper moral vision since the challenge we face today is that digital technologies are fundamentally altering the world, how we interact as a society and the range of our actions.

We need to move from existing ethics frameworks, which can be easily ignored or bypassed and amount to mere tick box exercise to a system based on virtue ethics that focuses on dispositions of character and cultivating morals – not just at the level of the individual but in companies as a whole.

What is taking place is a complete restructuring of the world according to an implicit philosophy, worldview and ideology, and the disruptive pace of digitalisation leaves little time for individuals and societies to adjust to the new conditions imposed by digital technologies.  As part of this restructuring, it is important to understand that features, objects and processes in the natural world are not translated into the digital domain as they are. Instead, they are reformatted to suit computational systems. In other words, the natural world objects and processes are forced to change or are replaced entirely with digital surrogates.

Let’s consider this restructuring in the context of digital devices that enable widespread tracking and surveillance of people going about their everyday life.  Given that digital devices are distinctly cultural products that radically alter private space, familial and work relations, the pace of life and the way society operates, one of the central issues in forming an ethical approach to technology is to understand the underlying ideologies, politics, worldviews and value systems that are inscribed in the very nature of digital technologies.

Given this situation, and speaking as a Muslim, what I would like to propose is that Muslims are in a unique position to understand the disruptive impact of digital technologies, to come up with ways that introduce new technologies in a more harmonious way, and even to act as a barometer for tracking the impacts of digital technology and the way they disrupt patterns of life and the ways in which society operates.

I say this because Muslims have a unique combination of ritual practices that link multiple orders of spatiality and temporality through practices that unite the body, community and natural phenomena in daily, weekly, monthly and yearly rhythms.  Digital technologies disrupt these patterns of life and sever the connections between space and time, body and environment and create new condition where the online and offline are seamlessly interwoven.

For example, consider the case of Muslim taxi drivers who make regular pauses in their trips at prayer times. It has been shown that they can be identified as Muslims with a high degree of accuracy from geolocation data that reveals the rhythm of daily ritual practices, this raises the issue of heightened surveillance, incorrect classification of Muslim populations and governmental and economic discrimination (Kammourieh et al. 2017, 47).

Second, consider the case of the connected workplace with continuous surveillance to gather productivity metrics of workers. For Muslims, work time may overlap with prayer times and going to pray may lead to a drop in productivity, ultimately putting practicing Muslims at a disadvantage, materially speaking. In this case, it may be argued that a workplace environment that pushes people so hard that it is difficult for Muslims to observe their prayers probably isn’t good for anyone else either. For example, the environment inside amazon warehouses has been criticised for this very reason where workers have been hesitant to even take toilet breaks to avoid drops in productivity.

Finally, Muslims are easy targets for discrimination in consumer data (such as purchasing halal meat (Zarsky 2014, 1395)) or in facial recognition systems (because of having distinctive appearances such as beards for men or the hijab for women). What I would argue is that data collection at a level of granularity that can identify Muslims directly or via proxies is highly problematic and such systems are likely to be capable of discriminating against a wide range of other groups with negative consequences.

One only has to consider the case of China where AI based surveillance effectively turns entire regions into detentions camps to understand the way digital technologies can be easily re-appropriated for malevolent ends and the urgency of the task of integrating a deeper moral and ethical vision into the task of developing and deploying new technologies.

In sum, it is beneficial to everyone for people of faith who use and develop digital technologies to think carefully about how their faith beliefs and practices are affected in the process of digitalisation and thus to participate in the ongoing societal discussion about the benefits and liabilities of such technology.

X