UMR HURTER reports on a recent webinar that sought to clarify the dangers – and benefits – of AI to the ummah.
THE onslaught of Artificial Intelligence (AI) on mankind has been years in the making, silently burrowing into the very foundations of human society. Science-fiction has become a reality and part of everyday life.
This message became very clear during a webinar held by Shaikh Dr Syed Mustafa Ali, Professor at the Open University in the UK on Thursday, July 20, 2023.
The webinar was part of the Tafakkur (the act of thinking, contemplating, reflecting) Series held by Strive UK. Dr Ali is a lecturer and convener of the Critical Information Studies (CIS) research group at the School of Computing and Communications.
The webinar titled ‘A Muslims Guide to AI’ sought to clarify the dangers AI poses to the Ummah. Interested parties from around the globe attended and were taken step-by-step on a journey on the why and the wherefore of AI.
Dr Ali explains: ‘I would want to stress the importance of thinking beyond the framework of benefit and harm and considering the matter in relation to power and Muslims’ position in the contemporary era (I perceive this as a shift from “izzat” to “zillat”). Additionally, it is important to understand that NO technologies are neutral; all have biases, values etc. embedded within them. Lastly, we need to consider how AI technology will be deployed against the Wretched of the Earth / Global South populations as well as the “Other” in the Global North.’
In his webinar, Dr Ali chose to focus on ChatGPT and its relation to Islam, as well as the danger of ChatGPT misquoting the Quran, providing inaccurate citations for Hadiths and Quranic verses, and declaring Haram things as Halal and Halal things as Haram.
He referred to a recent event where renowned Muslim scholar Shaikh Hamza Yusuf was invited to attend a symposium called The Rome Call for AI Ethics, an interfaith gathering under the auspices of the Vatican.
At this event, Shaikh Yusuf expressed a critical view on technology and progress in general. He also highlighted the distracting nature of technology and raised the important question of who benefits from this AI tech.
In contrast, the revered Imam Dr Omar Suleiman took a different approach, stating that AI was an inevitable development.
His approach to intervention was to draw on a spiritual framework, ethics and thinking in terms of principles, comparing the benefits and harms. In essence, Dr Suleiman referred to a cultural resistance to dependency, highlighting the dangers of becoming dependent on these technologies.
Mankind is already surrounded by artificial intelligence. From soap dispensers in public restrooms to controlling flight systems as we fly across continents and beyond, AI has propelled mankind forward with a significant push.
In short, artificial intelligence refers to a subset of algorithms that can glean patterns from vast amounts of data, enabling the performance of tasks that were previously exclusive to humans.
Artificial intelligence, in its purest form, aims to replicate human intelligence. It has developed exponentially, to the extent that self-awareness might very well become possible, with AI evolving into a sentient being capable of thoughts, emotions, and able to teach itself.
Currently, the growing fear and trepidation surrounding AI is based on undeniable facts and rooted in reality, backed by strong empirical evidence.
According to Dr Ali, the danger of AI lies in the possibility that society might eventually rely on it completely, not only for managing machine tasks but also as the sole trusted source of information. Additionally, AI could inadvertently make moral decisions for us. In short, it will render human cognitive reasoning obsolete.
‘The danger of AI,’ says Dr Ali, ‘is that it is built on information that is already out there, in many cases false information, especially when it comes to Islam. We have the scenario of garbage in and garbage out (GIGO). These technologies not only act as social mirrors but also as social amplifiers. They are not stable but rather in a constant state of flux. For example, ChatGPT 3 was more accurate in June 2022 than it was in March 2023. It seems to become more cautious over time. One should view AI as part of our societal infrastructure, as much as waterways and electrical grids, and not just as a hammer or drill that one picks up and drops again after use. It is here to stay.’
Dr Ali had the following to add: ‘Muslims are currently living in a state of zillah (an age of strong influence, so the way we need to think about things must be influenced, informed, and shaped by considerations of power first and foremost. We have witnessed significant advancements in reactive robotics, ranging from basic locomotion to linguistic intelligence, interpersonal intelligence, and even AI’s ability to negotiate and establish social relationships with people. Take, for instance, Tesla’s self-driving cars. The car should stop in front of a human crossing the road, but it should also stop in front of a stop sign. Since it lacks eyes, it operates based on pattern recognition detection. However, it is important not to believe the hype or be afraid of predictions that human intelligence will soon be surpassed by so-called machine intelligence. I do not believe that we will achieve full-blown sentience or anything close to our consciousness in a machine. Some Muslims may worry about whether a soul can exist in a machine, but for me, that is simply not possible.’
‘We must remember that the machines operate only on numbers, specifically two binary numbers: 1 and 0. The machine itself has no concept that it is working, no worldview, and it is simply performing correlations by matching inputs with outputs. Some applications of AI include self-driving cars, medical diagnosis, stock market trading, automatic language translation, speech recognition (SIRI), online fraud detection, and email spam and malware filtering.’
‘There are, however, problems with the Apps, some of which have still not been rectified. For instance, the Google Photo App recently misidentified an African couple as Gorillas. Another example involves two friends visiting a clothing store: the Hewlett-Packard Face Tracking Software successfully detected the white person’s facial features but failed to do the same for her black friend. This is a clear demonstration of how bias in the training data is replicated in the AI system itself. Another instance, specific to South Africa, is the Smart Soap Dispensers found in certain restrooms. When a white person puts their hand underneath the dispenser, the machine dispenses soap, but when a black person does the same, nothing happens. This is a deliberate design decision and another example of bias embedded in AI machines. Additionally, there is a recent case of bias in Face Recognition Software. An African female had to shine a light onto her face for the entire two days of her online exam so that the Face Recognition Software could recognise her and not flag her as a cheater. These examples clearly illustrate how structural injustices are being amplified and embedded into this new technology. It is crucial for us to remain aware and cautious.’
One of the cornerstones of Islam is our pursuit of knowledge from the cradle to the grave. As Muslims, we embrace science and have done so for centuries.
During the Dark Ages-Middle Ages in 15th-century Europe, citizens were burned at the stake for questioning the prevailing order of the day and refusing to live in blind faith.
Meanwhile, Middle Eastern scholars and scientists were actively exploring, investigating, and proclaiming numerous scientific breakthroughs, shining their beacon of light upon the world.
Practising ethical science leads to power, and when power is in the right hands, it can yield goodness. It is our responsibility to question and stay informed about how we wield this enormous power, lest it be misused.