📚 How I Read Thousands of Papers In the Age of AI

At a session for the Australian Public Service today, I gave a personal talk on how my reading life has transformed—both professionally and personally—through deep AI integration. With AI, I now digest over 2000 academic papers a year, share detailed notes with my team weekly, review papers for conferences and research , and still find ample time to enjoy 60+ novels annually, all read end-to-end in full immersion.

I thought it would be an intimate session with 50 people. Instead, over 1600 registered, with more than a thousand attending live!

This talk wasn’t about “productivity hacks.” It was about rethinking why and how we read in an age where AI can often summarise flawlessly, surface key sources, and even evaluate research. Once AI has “done the reading,” the real question becomes:
→ Do you trust it and use it directly?
→ Do you probe and verify it?
→ Or do you still choose to read the original—because some things must be experienced, not compressed?

I broke down reading into its primitive goals—discovering, selecting, extracting, understanding, evaluating, and experiencing—and explored how AI maps differently to each.

📄 I’ve uploaded my slides below. They include tools, workflows, and prompt examples from four use cases:
– Managing information overload through AI-curated feeds
– Digesting scientific papers with AI-assisted summaries
– Reviewing academic work with AI-generated critiques
– Reading fiction with AI as a conversational companion
I understand some of these examples could spark debate. :-}

Slides: https://www.dropbox.com/scl/fi/2b1hs3mvdygvx79wg58x3/Reading-in-the-Age-of-AI.pdf?rlkey=dj0upsiesbhzg78pwcqx1fp64&dl=0