Based on OnePath Network’s interview with Dr.Waleed Kadous, Chief Scientist in Silicon Valley
Artificial Intelligence, or AI, is now a fast developing reality that can replicate human voices and recreate the past using computer-generated images. Viral AI language systems such as ChatGPT are popular for the way they are able to produce customized, reasonably articulate responses to search queries.
Does this mean that humans will lose their jobs? Shall ‘intelligent’ robots replace humans? Will this lead to a rise in atheism, a reconstruction of religious thought and society?
Dr.Waleed Kadous, a computer scientist who is responsible for incorporating AI in Google, Uber and Google Maps helps us gain an understanding of AI and navigate the challenges that it poses to everyone, especially Muslims.
Dr.Waleed has been working for decades in software development. He explains AI as a programme based on learning from experience. Every time the algorithm sees a new example, it tries to see what it can be taught. Adaptation and modification was something that historically only humans did, but now computers can do too. Since AI is based on individual search questions and data input, it is personalized and therefore difficult to detect.
Challenges and fears:
- Do you trust the information from AI?
Dr.Waleed explains that he has been working with these programmes for a long time and he knows that they have a problem with what’s called ‘hallucination’. AI is not just ignorant but it ‘hallucinates’, i.e. it doesn’t know that it doesn’t know.
In other words, they can make up information and be 100 percent confident that its right- what is called ‘Jahl Murrakab’ in Arabic. Nothing stops it from making up a Hadith that doesn’t exist.
For example, Chatbots make things up and present it very confidently- they will never tell you that they aren’t sure. So be very cautious when using AI especially for religious search.
We need to understand these technologies and realize their limitations. We perceive them as somehow ‘magical’ and truthful and we ignore the historical biases that have existed in search engines. There is always a risk in these technologies of the ‘polarization effect’ if they are not used properly. So you don’t forbid the use of these technologies, but you do work on them to make them more consistent and effective for research purposes.
- So how do we combat false, new and artificial content?
Content is designed to press people’s buttons and this leads to confirmation bias. In Surah Al Hujurat, Allah (swt) tells us that when news comes to you, make sure to take time to confirm it. Verify it from different sources.
- Will AI affect social relations?
There is a possibility that it will affect social relations especially marriages. Marriage and birth rates are plummeting across Western societies and one of the main reasons for this is pornography. If AI can custom build a virtual companion, it will impact society by replacing real people and we are seeing its effects in an increase in social isolation.
- How biased are these AI Chatbots?
I once asked it to explain the Palestinian conflict to a ten year old but it gave me a dismissive response saying it was very complex. This is a highly illustrative example of bias in search engines that makes certain information difficult to search for or gives only one perspective.
Bias is very real because the humans developing these systems are themselves biased. There are two things to consider: one is the software (the algorithm) and then the other is the data that make the models which does the ‘thinking’. The software developer who is working on the algorithm and providing the data isn’t even aware of his bias so obviously it affects the way the search will be conducted. This has been proven comprehensively in the work of Timnit Gebro who used to work at Google and wrote a research paper on bias for which she was fired.
However, there are ways to reduce the bias. You can change the data that is used to train the model or in some cases you can change the ‘personality’ of the model to make it more Islamic. In my own research, the ‘personality’ change has gone surprisingly well.
- Does technology pose a risk of extinction?
Other causes of extinction such as climate change, the family unit being destroyed are more pervasive. Technology is still in its infancy. Whenever a new technology is developed, the creator scrambles to copyright it and the government steps in to build regulations so that it is controlled. So Sam of Open AI has an agenda to protect his business which is why he warns of job losses and the need for governement regulation.
- We’ve seen the impact of technology on our attention span. Will AI make us lazier or dumb?
Every time a technology is introduced, there is a fear that it will cause destruction of the human society. For example even when printing presses were introduced, people feared it would damage people’s ability to remember, but we’ve realized over time that they have had a positive impact.
Now we have to ask the question: have we gone too far? We live in such ease that it affects our value judgments. When knowledge was hard to obtain, we respected it more.
Earlier you would have to travel far and for a long time to learn a Hadith and now you can access it in 10 seconds, but has it made us better Muslims? There is a difference between knowledge and wisdom and this is the problem with technology as well. Sometimes when you make things easier and more accessible, you devalue it.
- Is there a trans-human agenda at play? A movement to merge humans and robots together?
There will always be people on the margins with crazy ideas such as a new species or who talk about ‘singularity,’ but it is not something I have seen.
- Could AI cast doubts on the awe inspiring nature of Allah (swt)?
If anything, it’s the opposite. There is a verse in the Quran about humanity creating a fly and we haven’t even been able to create a fly despite all these advanced algorithms, data and machines.
- Given that AI can create ‘matching images’ and replicate voices and even recreate the past to an extent, is it playing the role of God?
All these algorithms need something to get them started- it is simply clever ‘reassembling’. It cannot make something out of nothing so it makes us wonder how Allah (swt) does this. It should, in fact, make us more appreciative of Allah’s (swt) Powers as One who creates out of nothing.
- How far is AI from becoming sentient or conscious?
Scientists have struggled to define consciousness. The Quran says that knowledge of the Ruh is from the matters of my Lord. One of the earlier scientists said that asking if machines can think is like asking if submarines can swim- thinking is distinct from having a conscience or being ‘alive’ in that it can process and reassemble data fed into it to present different types of information.
- What are some potential uses of AI in the Islamic space that can have a positive impact?
One is Tarteel, an app that listens to your recitation and tells you when you make mistakes. Apps can create images that can cause us to reflect, so there are many exciting possibilities for every aspect of Islam.
Also, AI can capture semantic meanings and so can make Quran and Hadith search very effective and this could be very useful for scholars who can judge and integrate it carefully.
- How can we ensure that AI doesn’t ruin our relationship with Allah (swt)?
You can use AI to educate about Islam and develop effective Dawah material but you could also use it to make machines that kill and destroy. It depends on how we use it. It is like having access to books- you choose what to read and how to use that information.
- Where do you see AI taking humanity forward in the next 5-10 years?
All these technologies amplify trends that already exist in society so it is really up to us that we move in the right direction. I can see a world where learning about the world is going to be much accelerated, people feel more productive and meaningful at work because the menial tasks have been automated. But I can also see a world with more unemployment and more inequity in society and so as Muslims we have to ensure that the community and society are positively impacted.
- Should Muslims embrace AI?
A lesson from history has been that it has been a bad idea to lag behind. Take the example of the printing press that feared loss of memory and distraction. The same thing happened with satellites and television. If we have the right mentality and take the time to understand and channelize it, we can bring about more social good.
If we don’t have good Islamic models, people are going to use ChatGPT and so we need to be part of this narrative right from the beginning. We want this social change to play for us rather than against us.