Look far ahead
Since last month, I have been reading a rather scholarly book, The Book of Why, written by Israeli-American computer scientist Judea Pearl with Dana Mackenzie. The book deals with the often elusive relationship between cause and effect. Judea Pearl is the father of journalist Daniel Pearl, working for The Wall Street Journal, who was kidnapped and murdered by terrorists in Pakistan in January 2002.
One of the pioneers of Bayesian networks and the probabilistic approach to artificial intelligence, Judea Pearl, now 87 years old, argues in the book that while the traditional statistical approach was crucial, it has limitations and that understanding causality requires new frameworks and tools. He moots a “causal revolution” and seeks to establish causality on a firm scientific basis.
I was intrigued as this confirms the long-held statistical mantra, “correlation is not causation.” Looking for the “why” behind events is typical of Vedic inquisitiveness.
कोहं वा कुत आयातः किं कार्य तु मदीयकम् ।
कस्य पुत्रोऽहमुत्पन्नः केनैव निर्मितोऽधुना ॥
Who am I, where did I come from, and what is my job? Wohom am I brorn to, and by whom am I now created? (Shiva Purana Verse 2.1.7.8).
So, when I read about a professor of computer science and statistics and director of the Cognitive Systems Laboratory at UCLA talking about the pitfalls of relying solely on correlation, I felt proud of my civilisational heritage and saluted the ancient Indians for their profound insights.
The Book of Why delves into causal diagrams, powerful tools that visually represent causal relationships between variables. The book then gradually builds complexity, starting with everyday examples like diagnosing illness and tackling more challenging scenarios like policy evaluations and social science research.
No doubt, by understanding causality, we can predict, intervene, and ultimately learn from the world around us. This knowledge is crucial in various fields, from medicine and economics to artificial intelligence and law. However, the book’s impact extends beyond scientific inquiry. Causal thinking is highly beneficial in everyday life, enabling us to make better decisions, navigate complex situations, and understand the nuances of human behaviour.
I could not stop reimagining the tapestry of Indian history through counterfactuals—several “what ifs” emerged, presenting alternative possibilities about present-day India. While I acknowledge the dangers of historical determinism and oversimplification, counterfactuals can be powerful tools for critical reflection and engagement with the past.
What if the Mauryan empire hadn’t collapsed? What if a unified India, under Ashoka’s descendants, had created a robust bulwark against future invasions? What if the Gupta empire had embraced maritime trade more enthusiastically and integrated Southeast Asia with India? What if the Mughal empire had maintained religious tolerance? Akbar’s policies fostered pluralism until Aurangzeb turned fanatic. What if the British East India Company had never gained a foothold in India? What if India’s independence movement had taken a different path? What if the idea of a separate Muslim nation had not sprouted?
Looking more specifically, what if the Arab invasion of Sind by Mohd. bin Qasim had failed? What if Rana Sanga had defeated Babur? What if Dara Shukoh had succeeded Shah Jehan and not Aurangzeb? What if the Marathas had won against the English? What if the revolt of 1857 had succeeded? What if World War II had ended differently? What if Netaji Subhas Chandra Bose had returned? What if the partition had never happened? What if communist China had not swallowed Tibet? You and I would be living in a different country and, of course, a different world.
So, what next? Generative Artificial Intelligence is the new reality of our lives. It is foolish to imagine that it would be business as usual. Not only livelihoods but lives themselves are in for a transformation. Every sector is changing, and it is time to examine the counterfactuals not as some distant events in the future but in everyday life. What if I do not fix my unhealthy lifestyle, invest my savings, resist the corruption and criminality in public life, and provide children with relevant education? Reimagining the future with AI is indeed embracing AI now for good.
As AI evolves, which it is by the day, its ability to analyse vast data and predict trends opens a fascinating door to exploring “what ifs.” Counterfactuals, as thought experiments that ponder alternative outcomes, become increasingly relevant in navigating the complex implications of AI development. I can see five changes happening already. Their force and speed may differ in different societies, but no world region remains untouched.
1. Foresight and Risk Mitigation: Large datasets are available for every activity – from how many calories you eat, how many steps you walk, why you eat, to what is happening around you; identifying patterns is easy. Counterfactuals can act as a safety net, prompting us to consider alternative scenarios, saving us from imminent pitfalls, and helping us identify weaknesses and design more robust solutions.
2. Ethical Decision-making: AI algorithms learn from data, which can perpetuate biases and lead to discriminatory outcomes. The most significant bias in life is the feeling of “I-ness”. We embrace a lot of nonsense from our attachments and miss out on great benefits due to our aversions. Counterfactuals can play a crucial role in uncovering and mitigating these biases.
3. Policy and Planning: AI helps us understand complex systems and predict future trends, from climate change to economic forecasts. Most of the education today needs to be more relevant and functional. What human capital is expected from children spending their formative years in social media trivia and playing video games? What will be the new roles for those stuck up in repetitive jobs that are likely to be taken over by machines?
4. Creativity and Innovation: AI excels at automating routine tasks and analysing data, but it often needs help with genuine creativity. Counterfactuals can stimulate creative thinking by prompting us to consider alternative possibilities and break free from conventional constraints. Counterfactuals created by AI writing tools could prompt the tool to explore different narratives, plot twists, or character choices, ultimately leading to more original and engaging creative outputs.
5. Societal Understanding: As AI increasingly permeates our lives, understanding its societal impact becomes crucial. Counterfactuals can help us assess the potential consequences of income inequality. We ought to be sensitive and kind to the poor and disadvantaged. No person is an island anymore; this must sink in no matter what we do.
Counterfactuals are not prophecies, and making deterministic claims about the future is foolish. However, counterfactuals can be a powerful tool when used critically and responsibly. Your future will depend on what you are doing today. Internalise this shloka in your life:
धर्म चरत, माऽधर्मंम्, सत्यं वदत, नानृतम् ।
दीर्घं पश्यत, मा हृस्वं, परं पश्यत माऽपरम् ॥
Practice righteousness and avoid doing that which is incorrect, speak the truth and refrain from lies, do not live with a narrow vision, look around and be farsighted, and examine yourself and don’t keep looking at others. (Mahabharata 3. 163. 31)
By encouraging us to question assumptions, explore alternatives, and anticipate potential consequences, our daily lives will help us shape an extraordinary future while AI benefits humanity. Know AI as the collected and curated natural intelligence of the entire humanity at your service.
MORE FROM THE BLOG
Take the Bull by the Horns
I am a mechanical engineer who worked for 15 years at the Defence Research & Development Laboratory (DRDL) in Hyderabad. There, I developed the Trishul and Akash missile airframes and Titanium Airbottles…
Mythology: Imagination or Coded Secrets?
Myths often reflect the creativity and storytelling traditions of cultures. They can be seen as symbolic narratives that explain natural phenomena, human behaviour and societal values. These stories often incorporate fantastical elements and characters…
Whose Truth and By what Facts do I Live?
I exist. My family, friends, and people I wouldn’t say I like all exist. I face situations both comfortably and disquietly. I feel happy and sad, relaxed and anxious. Like a flowing river, the stream of thought in my mind sometimes flows smoothly…
Interesting insight into causality, Prof Tiwariji !
Your concept of AI and its societal impact is outstanding !!
In his famous book ‘How to Win Friends and Influence People,’ Dale Carnegie writes, ‘By becoming interested in the cause, we are less likely to dislike the effect.’
Reasoning about causes and effects is tricky, and I believe God is the only judge of that. We are already hard-pressed to establish a relationship between such a pronounced effect as a charred tree and the lightning bolt that set fire to it, so to trace sometimes endless chains of causes and effects seems to me as foolish as trying to build a tower that will touch the sky.
Thank you, Sir, for encouraging us to reflect upon such topics.
Thank you, Sir, for introducing us to a new author, a new book, a new mental model and explaining the AI subtleties. In generative AI, we now have a tool to reinvent and change the Business as Usual. It is timely as it unfolds when climate change’s wicked problem seems unstoppable. Summarising AI as a curated natural intelligence is beautiful.
Lovely reflection! In science, the cause explains why something happens. The effect is the description of what happened. Children are born because their parents mate. Plants grow as a result of the sun’s heat and light. Bears hibernate to conserve energy and won’t require food during winter when food is scarce. The moon has a gravitational pull; consequently, the oceans have tides. Destructive lifestyles cause chronic diseases. There is growing evidence that pollution is causing climate change.
Using AI, deep and hidden connections can be found. Like distant landmarks you are approaching, cause and effect begin aligning and drawing closer together, creating the phenomenon. Experiences too indefinite of outline in themselves to be recognised for themselves connect and are identified as a larger shape. And suddenly, light is thrown back, as when your train makes a curve, showing that there has been a mountain of meaning rising behind you on the way you’ve come, is rising there still, proven now through retrospect.
Super interesting!!
I recently read an article on Generative AI becoming conscious. To be exact – What if generative AI becomes conscious? Our students are working on a project on the same. The real catch, however, would be how to define consciousness. Psychologists define consciousness under three levels – unconsciousness, preconsciousness and consciousness. However, philosophers say it varies from self-awareness and survival instinct to societal contribution. The danger is that a conscious AI might have contrived inducements such as gaining more ‘likes’ on any social media platform.
Whatever it may be – GENERATIVE AI IS HERE TO STAY.
Dear Sir, the cause-and-effect mental model is a valuable tool for making better by identifying the root cause of a problem. This mental model assumes that every effect has a reason and every cause has an impact. By identifying the cause-and-effect relationships, we can better understand the factors contributing to a particular outcome and make informed decisions to address the problem. Understanding this mental model can help individuals and organisations improve decision-making processes and outcomes.
But while the cause-and-effect mental model is valuable, there are some challenges. One of the most common pitfalls is making assumptions or having biases. Then, there are issues with a lack of data or incomplete information, complex or interconnected causes, overlooking secondary or indirect causes, and failure to follow through on an action plan. But living mindfully with cause-and-effect awareness is undoubtedly beneficial. Thank you for writing on such an essential aspect of life.
This Blog brings up some big questions. If history could be different because of small changes, does that mean our choices don’t matter? How do we really know for sure what causes something to happen? Thinking about “what ifs” can be helpful, but getting stuck in them might stop us from acting in the present. As machines get smarter, can we teach them right from wrong, especially when things could have gone many ways? It’s interesting how people throughout history, like those in ancient India, have always wondered about cause and effect – it seems everyone wants to know “why”!
Sir, this verse from ‘Aparokshanubhuti’, a small booklet by Adi Shankaracharya, is an excellent mantra for self-realisation.
कार्ये कारणताता कारणे न हि कार्यता।
कारणत्वं स्वतो गच्छेकार्यभावे विचारतः॥
Causation arises through action only. It is not right to blame causality for action. The actions done in the past are the reason for everything good and evil in the present.
Cause and effect together make phenomena. There is a reason behind everything that exists. For example, overeating fast food without physical activity leads to weight gain. Eating without physical activity is the “cause”, and weight gain is the “effect.” But is it so? Why is one prone to eat more? Why crave binge eating? Can I consider myself an extension of my parents, whose DNA is my existential core?
Another famous example in the discussion of cause and effect is that of smoking and lung cancer. A question that has surfaced in cancer research in the past several decades is, What is the impact of smoking on an individual’s health? Does smoking cause lung cancer? It turned out that the tar that enters the lungs causes cancer. But that doesn’t absolve cigarette smoking. How does tar reach the lungs? This blog offers very interesting insight indeed!
Very interesting takes, Arunji. You set a grand narrative on how AI in general and counterfactuals, in particular, can be leveraged positively for the greater good.
I also loved this beautiful articulation – “Know AI as the collected and curated natural intelligence of the entire humanity at your service”.
Recently a friend asked for some paintings on 64 kalas with Sri Krishna’s theme, I tried creating few of them with Dally 3.0 giving some very basic prompts and the outcomes were quite impressive.. The image it created for Sri Krishna engaging in the Art of carpentry with Yamuna in the background was so beautiful that we all loved it instantly.
Thank you for sharing Arun ji. ‘The book of why’ is one of my favorite books. I loved how you have connected it back to the wisdom of the ages.
Sir, how you connected probabilistic theories of correlation and causation to what-ifs from different parts of the world and now look at how AI is transforming our lives and how to look far ahead is fantastic. Also the last two paragraphs remains same whatever happens around us….Keep sharing your thoughts.
Prof Tiwari causality vs coincidence is very interesting thought in medicine as well. The world of medicine is now fully immersed in the new kind of AI and its applications and challenges.
Dear Sir, Thank you for sharing your thoughts on AI, I am also currently working on an AI project with HCF, Australia (The Hospitals Contribution Fund), where we are creating a virtual live person to assist PHI members when they will visit the HCF website for their queries. The virtual person with the help of AI will be able to answer the queries from the members and guide them to their need. It’s going to reduce the online customer care calls to a great extent.
Undoubtedly, AI will be an essential tool in the hand of humanity the way people and business interact in the future. It will increase the job opportunity in technical and operation filed as there will be requirement for more supervision than to involve in doing the work at workplace. Warm Regards,
This just sums it: questioning assumptions, exploring alternatives, and anticipating potential consequences. What can we possibly not do or achieve through AI and counterfactuals? The power to redefine and offer alternative pathways to every day challenges is being availed at hardly any cost say for a curious mind.
The area of Explainability in AI prediction models is a key focus in the industry. We are actively looking at promising approaches to improve on this. Another area of concern for the future computing requirements of AI is power consumption. Sam Altman, CEO of OpenAI, was rumored to have been talking to UAE investors for up to $5-7 trillion (not a typo) of funding – supposedly to build a network of advanced chip fabs, datacenters and renewable energy plants. Microsoft just announced a 1.58 bit technology that works well for LLMs which could dramatically reduce the compute requirements for LLMs. And we’re working on quasi-adiabatic logic technology which dramatically reduces logic power. Exciting times ahead for humanity with Explainable AI — providing the answers to … WHY?
As someone currently immersed in Epidemiology and Biostatistics, I couldn’t agree more with Professor Judea Pearl. I am very much looking forward to the development of advanced methods which can predict causality more accurately. I also agree with you on the various changes counterfactuals can help us with. I find its usefulness in policy and planning most intriguing, and I will try to implement it personally to help myself make ethical decisions.
Dear Prof
Beautiful piece! Thank you for sharing.
Thank you for your reflection Prof Tiwari, I agree: ‘As AI increasingly permeates our lives, understanding its societal impact becomes crucial. Counterfactuals can help us assess the potential consequences of income inequality. We ought to be sensitive and kind to the poor and disadvantaged. No person is an island anymore; this must sink in no matter what we do.’
In the New Testament, we are reminded by the Apostle James that Puredefiled before God and the Father is this: To visit the fatherless and widows in their affliction religion and un and to keep himself unspotted from the world. James 1:27 – KJV
May this admonition find reception in our lives, everyday, everywhere!