On a quiet weekday afternoon, I decided to take a break from my desk and listen to a song. Not on YouTube, Spotify, or Alexa, but ChatGPT. I hear it can do “everything” so, why not? ‘Hey ChatGPT, please sing me a song’, I asked. A few verses of an ‘uplifting and catchy’ song quickly appeared on screen and I was asked to ‘imagine it with a gentle acoustic guitar rhythm’. Not quite what I had in mind.
Well, not music, but it could definitely make me some food. ‘Hey, ChatGPT, please make me some ramen’, and quicker than two-minute noodles came a recipe. I insisted; ‘Make it for me’. ‘I wish I could… but until teleporting ramen becomes a thing, here are more suggestions…’ came the reply.
Nevermind, ChatGPT. You can create resumes and cook up stories, but there are things you will never be able to do. Experts across the field of fashion, art, writing, design, and more, tell us why.
Gresham Fernandes
| Photo Credit:
Illustration by Saai
Gresham Fernandes, chef-partner at Bandra Born
As someone who’s obsessed with the rawness of art — whether it’s the hand that draws a line or the mind that layers a beat — I find myself caught in this odd, uncomfortable in-between.
In a kitchen, it’s not just about flavour combinations or plating techniques. It’s about the instinct to add a squeeze of lime because your grandmother used to. It’s the hesitation before you plate, the tiny adjustments you make after tasting something for the 100th time. It’s the way a smell can stop you mid-step and pull you into a childhood moment — and how that moment becomes a recipe. AI can suggest ingredients, and it can follow trends. But AI doesn’t know why you reach for mustard seeds instead of cumin on that particular day. It doesn’t know what grief tastes like in a broth, or how joy needs just a touch more salt.
I draw. I DJ. I cook. These aren’t just things I do — they’re how I see the world. And I can tell you, with full certainty, that the one thing AI will never replicate is intent. That quiet chaos behind every brushstroke, or the energy of a live set that shifts with the crowd, or the way a dish evolves from a memory into a plate — that’s deeply human. AI can mimic the output. It can even make you feel things. But it cannot feel.
AI is a business, and it cannot be an artist, says Appupen
| Photo Credit:
SvetaZi
George Mathen aka Appupen, illustrator
The copyright conversation is as old as AI. I have detailed it in Dream Machine, the graphic novel on AI I brought out with entrepreneur Laurent Daudet. (‘With Dream Machine, my focus was to impress upon the reader that the hands behind any such technology are rarely without bias and agenda. My attempt was to look at artificial intelligence without the sparkling halo and hype. ‘AI is just a tool, like an engine. We don’t have to relive one more disaster to prove that reckless profiteering is not great for us. It’s bad for the planet and all life on it’, he writes on dreammachine.in.
AI cannot create, it reproduces information from the training data. AI is a business, and it cannot be an artist. It will churn out faster and cheaper for mass consumption.
Also, AI can’t seem to be accountable thanks to the businesses that run them. If I make a spoof poster on Amul butter, I’ll get a legal notice immediately, be asked to take it down, and threatened all the way. I can’t believe that Studio Ghibli can’t get back at Open AI, so it seems AI cannot be caught. Or are they working together for mutual benefit?
Aarushi Gupta
| Photo Credit:
Illustration by Saai
Aarushi Gupta, senior research manager, Digital Futures Lab
Given the hype surrounding AI, it seems like it can pretty much do everything. At least, that is what the popular discourse will have you believe. But a closer look at a generative AI model — through regular use and a bit of stress-testing — reveals not a ‘superintelligent’ entity but a pretty ‘normal’ technology, as computer scientists Arvind Narayanan and Sayash Kapoor recently put it. These models might be really good at interpreting natural language or generating Miyazaki-esque images, but they do not do well on problems that require contextual nuance and analytical sophistication.
There are many aspects where Generative AI falls short when it comes to the field of policy research. It convincingly hallucinates, fabricating well-formatted citations and sources of the text it has generated. Such instances can profoundly undermine the quality of policy research outputs if Gen AI tools are used uncritically for research and writing. AI cannot generate new ideas: generative AI models are trained to predict the next word based on statistical probability. While the text they generate may appear convincingly human and coherent, it is ultimately a sequence of probabilities masked as a language we recognise. These models can mimic us, but cannot think like us. Each of us brings rich, lived knowledge, contextual awareness, and a deep understanding of the world — qualities that are essential to meaningful policy work.
AI does not truly understand the words it produces; it lacks intent, insight, the ability to grasp nuance, and most importantly, in the context of policy research, the ability to empathise.
Sanjay Garg
| Photo Credit:
Illustration by Saai
Sanjay Garg, designer
AI cannot create the basic realities, and necessities of our existence. It cannot eat, drink, have sex… cuddle, or cry or feel. It can only do things based on what’s human: imitating. But human evolution is a living phenomenon, an ongoing and growing process.
Though AI may be a part of the future, it can’t predict the future. Nor can it be the only future. It’s funny, from Early Man and the Stone Age to the age of AI, the basic things have to be done (and can only be done) by humans.
For me, in my field, it’s amazing. The more high-tech the world goes, the more special my work is considered to be. Hand-drawn, handcrafted, drawing on centuries of human knowledge passed down. My value is up because I am not an AI.
India is a land of stories, and oral storytelling is a major part of our culture. It doesn’t come from one source. It’s not so evenly distributed and it’s difficult to organise it so neatly, but it is still sustained. There are thousands of practices, religions, income groups and age groups, languages, skill sets, customs, tribes and cultures and experiences: all of these inform what we know, and this plurality lends itself to our individuality.
So much of this is undocumented, that even to record all of it, it will take a long time. This is what’s nice about India. Its strength and its weakness, this undocumented, chaotic repository of knowledge. That’s exactly what AI can’t know, it can only draw from what is documented and online. There’s emotion, unique self expression and connection. Our idea of what’s luxury and what’s exclusive? That’s something that only the human hand can do.
Anu Menon
| Photo Credit:
Illustration by Saai
Anu Menon, actor-comedian
No AI robot will have the obscene level of patience required to deal with a pre-teen! My labour was easier than getting my son prepared for his exams. Forget just hormonal mood swings… Gen Z/ Gen Alpha lingo… human parents have to equip themselves to raise children in an AI world.
As someone who has never used AI for writing, there are so many things it cannot do. For starters, a comedian can laugh at his/her joke. A comic can take down a heckler best because when it is happening mid-show, you don’t have time to ask Chat GPT for the best response. Also, when a comic is talking about a personal experience, only he/she can capture it best because either they or someone they know has gone through it. AI can never capture or express that experience because it is personal.
Art forms are the final frontier that will fall to AI. There isn’t a fool-proof cheat code for individual artistic expression. I’m sure AI can give you joke ideas, but I’m old school, and believe that a joke has to be in your voice. I don’t think we should invest in our own demise. Can we have more chauffer robots and house cleaning robots? That’s the need of the hour.
Ananya Singhal
| Photo Credit:
Illustration by Saai
Ananya Singhal, architect and co-founder, Studio Saar
The idea that AI will replace human work is disingenuous and absurd. How humans survive and work has constantly changed over time, because we began to use tools that made life easier. AI is another tool that will do just this. We can’t talk about never, because AI is a part of the evolution of human invention.
Having said that, what it can’t do in the foreseeable future, is quite simple. AI is not able to have the ‘eureka’ moment, the sudden flash of insight that comes from the creative mind making connections that were unknown. As a tool, it enables us to find all the threads and connections, it doesn’t and probably will not find that unique connection that is the true problem solving.
For that, you need human beings. You may find pathways and AI will present the pathways to architects. How do we navigate them in our own ways? That is our challenge. Much like Google Maps shows us the ‘best’ way through a crowded city, and yet we use our own judgement.
A key part of architecture is communication. While AI may be getting better at it, architecture itself is changing. Today, we engage the community, we work with the future users of our spaces to understand their needs and to help them in their journey to understand the new place we will be making together. This is a truly human activity, one that will remain human.
Yatin Pandya
| Photo Credit:
Illustration by Saai
Yatin Pandya, principal architect, Footprints E.A.R.T.H.
Design is a dialogue between the creator and perceiver/user. The designer encodes the clues and the perceiver decodes the same. For this bond to be effective it is essential to understand the subtleties and spontaneity of the milieu. Like a doctor, it is imperative for the creator to have unique, personal and one-to-one bond and it is in this process that AI generated products fall short of.
The most critical shortcoming of AI is in developing an emotional and experiential connection to spaces. Spaces intuitively and intimately connect to a person and generate a sense of belonging. Thus, it is a spatial journey with designed encounters. The essence is in mitigating the meanders and experiencing the sense of surprises and muses. This needs a very humane conception of spaces which is far deeper than the form and geometry. Spontaneous adjustments, understanding the tone rather than spoken words, reading unwritten words is what dialogue is all about: Intuitive iterations.
Bijal Vachharajani
| Photo Credit:
Illustration by Saai
Bijal Vachharajani, author
AI can never recreate the joy of a child reading a book. When I see children lost in a book, responding to each page turn, I know that is what human-crafted work does. It offers a space to respond to the story written and illustrated, edited and art directed, copy edited and proofread by people, who each responded to that narrative in their own individual way.
The spark of an idea, watching it bloom into a full fledged book, is something an AI bot cannot do. That first glimmer of a thought, the building of it, the frustration, the ‘aha!’ moments, and then watching it all come together. Sometimes to fall apart, sometimes to become something new.
Technology can never build a connection with a character. We find ourselves in books, in its people, its animals, its places.
