Latest news with #KnightCentre

Hindustan Times

Book Box: How to cope with AI anxiety

Dear Reader, Empire of AI ticks all the boxes. These days, I see AI writing everywhere—on LinkedIn, in text messages from colleagues, and even in substack newsletters. There's something about these polished pieces of prose, glib and formulaic, with their idiosyncratic sentence structures and excessive dashes, that end up depressing me. Many of my writer friends won't touch AI. 'We can write just fine without it,' they say. But I can't stay away. I face my AI anxiety by finding out what this new beast is. As a teacher of management, and as someone who has pivoted careers three times already, I feel compelled to keep up with the times. AI making all writers redundant I sign up for 'Prompt Engineering 101 for Journalists' conducted by the non-profit Knight Centre. It teaches me how to prompt AI to 'red-team' my writing—to critique flaws rather than default to dishing out praise. And to watch out for AI 'hallucinations' like made-up names of books and fake quotations falsely attributed to real people. I stay conflicted: is it okay to use large language models that ride on the backs of writers and artists, that have learned by scraping creative works with no regard for privacy or copyright? And what about the environmental toll—the depredations on water and energy that the data centres inflict, especially in developing countries? I look for my answers in books about AI. Supremacy: AI, ChatGPT and the Race That Will Change the World is a 2024 book by Parmy Olson that won the Financial Times Business Book of the Year. It takes me close to AI stars like Sam Altman of OpenAI and Dennis Hassabis of DeepMind, as well as to the dangers of decision-making being left to a tiny elite. But it leaves me wanting more. A friend recommends The Worlds I See: Curiosity, Exploration, and Discovery at the Dawn of AI, a memoir by Chinese-American scientist Dr. Fei-Fei Li. It begins with an exciting chapter, with Fei-Fei traveling from her West Coast Google office to Washington, D.C., to testify to a US Senate committee on the direction of AI research. For many pages I am enraptured, reading Fei-Fei's family history, how she helped her parents run their dry cleaning store while studying at Princeton and then working at Stanford. Li has been a pioneer in AI image recognition with her ImageNet project, and this makes for fascinating reading. The book veers between scientific excitement, and apprehension at where AI research is going, and confirms my unease over the economic and existential implications of this new technology. Then I discover Empire of AI by Karen Hao. From the very first page, I am highlighting lines, drawn in by Hao's historical analysis of AI research over the years, everything from the 'AI winter' to the dispute between two schools of AI research—the symbolists and the connectionists. Empire of AI ticks all the boxes. It is rich in history and human detail, demystifying core concepts like deep learning and neural networks. Hao gives us the stars like Geoffrey Hinton, Ilya Sutskever, Greg Brockman—and also the workers, the data labelers and content moderators like Mophat Okinyi, Oskarina Fuentes Anaya, and shows the havoc that AI jobs have brought to their lives, as they are forced to deal with explicit sexual content and violent images and to perform AI training tasks for a pittance. For me, the most moving part is the story of Sam Altman's sister Annie Altman, who turned to sex work, having suffered huge health challenges and trapped in severe financial duress, against Sam's lifestyle featuring multimillion-dollar homes and luxury cars. 'Annie's story also complicates the grand narrative that Sam and other OpenAI executives have painted of AI ushering in a world of abundance. Altman has said that he expects AI to end poverty... And yet, against the reality of the lives of the workers in Kenya, activists in Chile, and Altman's own sister's experience bearing the brunt of all of these problems, those dreams ring hollow', says Hao. I put aside Empire of AI to go back to my day. I know it's ironic and it feels very meta, but after writing this, I ask Deepseek to design a brief depicting a writer dealing with the good and bad sides of AI, and then I use that output to ask Gemini to design the illustrations for me. AI, the perfect productivity tool ? That evening as I walk down towards the market to buy a AI-recommended geyser, I find myself grateful for Karen Hao's book. Because if AI's future is being written by the Altmans and the Musks of the world, excluding large sections of the world, there are things we can do to participate. Reading books like Hao's pushes us to pay attention—to the workers behind the algorithms, to the biases in the data, to the futures we're building one query at a time. Books like these arm us to fight back - to push for policy changes, demand transparency in training data, and support ethical AI movements. So yes, I will use AI. But I'll also keep reading and buying subscriptions to real writers and real news outlets, because the best defence against a dystopian future is to dream of a better one, and then to fight for it. What about you, dear Reader? Do you find AI more anxiety-inducing or enabling? Or a complex mix of both? And can you suggest any other such books on AI that we can add to this vital reading list? (Sonya Dutta Choudhury is a Mumbai-based journalist and the founder of Sonya's Book Box, a bespoke book service. Each week, she brings you specially curated books to give you an immersive understanding of people and places. If you have any reading recommendations or reading dilemmas, write to her at sonyasbookbox@