I have a Claude subscription. The package, in both its affordability and its efficiency leaves me suspended between two states of mind: a smug complacency regarding my professional future, and a creeping fear of redundancy. Between the comfort of the first and the anxiety of the second is an uncertain gap, which much of the world now inhabits, owing to rapid advances in Artificial Intelligence (“AI’). This liminal space will soon become the preserve of our children and grandchildren. How should the coming generation brace themselves for and in a world where Information Technology brings not just change but upheaval? Is there a way of ‘future proofing’ a future that is unpredictable and invisible?
Writer, economist, and also a researcher at the Institute for Ethics in AI at Oxford University Daniel Susskind addresses this existential question in his essential book, What Should My Children Do? Warning about engaging in futile attempts at ‘futureproofing’, Susskind contends that the most practical way to progress and prosper in a world getting ever closer to Artificial General Intelligence (“AGI”) is to adopt what he calls, “the teach both and test both” method: teach children to use AI, teach them to work without AI, and test both.
The “teach both and test both” method is neither new nor novel. The humble calculator arrived in classrooms in the mid-1970s and immediately triggered a crisis in the teaching of mathematics. Students reaching for the gadget was so common that in 1979, the British Government asked Wilfred Cockcroft, a now largely forgotten mathematician from the University of Hull, to produce a comprehensive report. Cockcroft’s solution was simplicity itself – teach students with calculators, teach students without calculators, and test both.
I have benefitted from the “teach both, test both” method myself. While reading a dense work of historical fiction on the famed Chola empire, written in my mother tongue, Tamil, by the late Balakumaran, I got impossibly stuck on a passage involving the rescue of a bullock cart caught in the swelling waters of the river, Cauvery. Giving up repeated attempts to comprehend the rescue scheme, I photographed the relevant paragraphs and uploaded them to Claude. I asked Claude for a concise infographic of the plan. I then “stress tested” the infographic against the basic laws of Physics that it elucidated, looking for inconsistencies. I found none. For the intrigued here is the image:

While largely in agreement with Susskind’s “teach both, test both” method, I would be wary of the “teach both” part of the approach, where over enthusiasm risks defeating its very purpose. Stringent, uncompromising lines must be drawn where teaching with AI ends and teaching without AI begins. Without such clear delineation, the “test both” aspect stands severely and unfortunately diluted.
If I had to choose one enduring ‘takeaway’ from the book, it would undoubtedly be the premise and backbone of its concluding chapter imaginatively and aptly titled, “Focus on Who You Are, Not What You Do.” Susskind urges us to teach our children and grandchildren the moral character that would enable co-operating with AGI instead of competing with it. There would be jobs that would still remain the preserve of human beings, especially the ones demanding a great degree of empathy. A patient in palliative care, I suspect, would feel more secure holding the hand of a caring nurse in flesh and blood rather than preferring the company of a humanoid howsoever soothing its words howsoever precise its care.
Yet, while acknowledging the inevitability of treating AGI as a trusted companion in future, we must ponder the impact of such a partnership, without guardrails setting out principles of ethics and security. The last thing future generations need is an AGI whose goals drift dangerously away from those of the humans employing it. That would be a path to perdition.
Meanwhile, I will continue to bask in the artificial security my subscription provides, while tightening my grip on the esoterica of international taxation.