
Mr. Jaspreet Bindra
Founder, AI&Beyond
“Intelligence becomes free — on tap.”The core idea
Session recap
Direction sense: what AI is actually doing to our world
Jaspreet Bindra opened Day 2 by refusing to treat AI as a software topic. He argued that every leap in human history came from a set of tools that made something abundant — agriculture made food free, the industrial revolution made labour abundant, the internet made information free. Artificial intelligence, he said, is the revolution in which **intelligence itself becomes free, on tap**. To make the point concrete, he showed a full enquiry-to-advocacy map of a school's business that he had built in minutes, knowing nothing about running a school.
That demonstration carried a second lesson he kept returning to. Generating the map took roughly twenty-five seconds; the prompt took four minutes; verifying it took six. "It took me longer verifying than creating" — and that, he told the chairpersons, is exactly the discipline AI demands of a leader. Speed is free now. Judgement is not.
He then walked the room through the shape of the field: AI is seventy years old, coined in 1956, but most of us met it on 30 November 2022 when ChatGPT arrived. Before that we had predictive AI, quietly forecasting sales, weather and traffic. Generative AI generates; predictive AI predicts — different animals for different jobs. Beyond generative sits agentic AI, systems that do work rather than answer questions, and beyond that the physical AI that Jensen Huang calls the biggest wave still to come, arriving first as autonomous vehicles rather than humanoid robots.
“Intelligence becomes free. Freely available, on tap. That is what this revolution actually is.”
The 2017 defeat of Go world champion Lee Sedol by AlphaGo was, for him, the turning point: Go is intuitive and probabilistic, not deterministic like chess, and its fall showed that a non-human intelligence could be better than us at genuinely human things.
Then came the part that landed hardest with a room of school owners. For our generation, the passport to success was English. For our children's generation, it was coding. AI collapses both: every prompt is code, written in your own language instead of the machine's. Potentially 140 crore Indians can now instruct a machine. The interface is collapsing too — DOS gave way to graphical screens, screens gave way to apps, and apps, he predicted, are going to die in favour of voice. "One year from now we'll look back and say — I used to type." Search follows the same path: we never wanted ten blue links, we wanted answers, so Google becomes an answer engine and the ad-funded internet underneath it starts to wobble.
On hallucination, he was unsentimental. AI gets things wrong, much like people do: "Tell me one employee of yours who gives you the right answer all the time, and I want to hire that fellow." The answer is not distrust, it is verification — the six minutes, every time.
Editorial summary compiled by the LDP team — not a verbatim transcript. Spotted an inaccuracy? Let us know.
Watch the session — AI for Edupreneurs, Day 2 of INSIGNIA 2026.
Key takeaways
- 1
Free intelligence changes the job, not just the tool
Treat AI as a general purpose technology on the scale of electricity or the internet, not as one more piece of school software. Ask what becomes possible when analysis and drafting cost almost nothing.
- 2
Spend most of your time verifying
Creation is seconds; verification is where the value and the risk sit. Build a habit — and a rule for your teams — that nothing AI produced goes out without a human check against what you know.
- 3
Know which AI you are using
Predictive AI forecasts (admissions, attrition, results). Generative AI creates (letters, plans, material). Agentic AI does multi-step work for you. Choosing the wrong one is the most common mistake.
- 4
Language is the new programming language
Anyone in your school who can read and write can now instruct a machine. That widens who can build with AI far beyond your IT team — and makes prompt clarity a leadership skill.
- 5
Move to voice earlier than feels comfortable
Typing is already the slow path. Voice interfaces are the fastest growing way people work with AI, and the ones that fit a principal's or chairperson's day best.
- 6
Plan for the answer engine
As search becomes answers, discovery of your school shifts from ranking on Google to being represented accurately inside AI answers. Start thinking about how AI describes your institution.
What he spoke on
- AI as a general purpose technology — and why intelligence becoming free is the real story
- A short history: 1956, predictive AI, ChatGPT, agentic AI, physical AI
- AlphaGo and the moment machines beat us at an intuitive game
- English as the new coding, and vibe coding in natural language
- The death of the app and the rise of voice as the interface
- Search to answers — and what that does to the internet's business model
- Hallucination, verification and the leader's judgement
Q&A captured
Q. AI keeps citing sites that don't look legitimate. Can we trust what it says?
It hallucinates — and so do plenty of legitimate sources. The answer is not to avoid AI but to verify. Spend more time checking than generating, exactly as with a new employee whose work you would review before it went out.
Q. Is AI really new? It feels like it appeared out of nowhere.
The term was coined in 1956. What changed on 30 November 2022 was that generative AI became something anybody could talk to. Seventy years of work reached a usable interface.
Q. What is an AI agent, in plain terms?
A tool helps you do work. An agent does the work for you — it plans, takes steps and comes back with the outcome rather than a suggestion.
