How to Think About AI: A Guide for the Perplexed by Richard Susskind
Published in May 2025
In the AI era, recognized expertise will matter more than ever. As it becomes more difficult to distinguish between human and AI writing, we will increasingly depend on our evaluation of the author’s credibility in deciding what to read. Call it the AI Matthew Effect: Those with the most existing cultural capital will disproportionately benefit from rapidly improving AI, while authors newly building their reputations in the age of AI will struggle.
The books we choose to read on AI are not immune and may be even more susceptible to our prior author judgments. So much is coming out about AI nowadays, and AI technology is evolving so rapidly, that the AI signal-to-noise ratio makes choosing our AI books super difficult. The reason I read Richard Susskind’s 2025 book, How to Think About AI, is that I’m a big fan of books by his son, economist Daniel Susskind, including A World Without Work (2020) and Growth: A History and a Reckoning (2024), as well as their co-written book The Future of the Professions (2016). Because I know and like Daniel Susskind’s work, I plan to read his upcoming book, What Should My Children Do? A Human’s Guide to the Age of AI, when it is published in November.
While How to Think About AI is not co-written with his son, Richard Susskind frequently references Daniel’s writings, making it clear that father and son are AI thought partners. Richard Susskind has been thinking and writing about AI since well before ChatGPT launched in November 2022. His writing on technology and the law goes all the way back to the late 1980s, and since the 1990s he has been working and writing at the intersection of AI and the legal profession.
This early work in AI set Susskind up well to think about generative AI, but also (as he freely admits in the book) created blind spots in how quickly AI would advance. It is worth remembering that prior to the release of ChatGPT, most “AI experts” focused on narrow AI use cases and were largely dismissive of the idea that AI would threaten professional knowledge work in the near future.
What makes How to Think About AI useful, and why I recommend the book as a good foundational text to get us all speaking the same language about AI on our campuses, is Susskind’s call for us to think today about what it will mean when artificial general intelligence arrives tomorrow. By tomorrow, I mean that we don’t and can’t know the timeline of when an AI that is as capable as humans across a wide variety of tasks will arrive. Susskind makes the case that AGI-adjacent capabilities are close at hand and will be hugely disruptive to our organizations and societies when they do emerge, so now is the time to figure out how we should respond to that coming eventuality.
A strength and a weakness of How to Think About AI is that Susskind broadens the AI lens to consider the implications of the technology beyond the coming changes in professional work. I enjoyed and learned some things from the discussion of societywide AI risks, debates about AI consciousness and AI ethics. Susskind is persuasive in arguing that AI technology is too important to leave up to AI companies in deciding how to utilize and regulate.
Where I wish Susskind had spent more time in How to Think About AI is how we should think about AI in our own work, and the work of our colleagues. This lack of consideration for how AI is already changing the day-to-day work of knowledge professionals (including most of us who work at universities) may indicate how quickly AI technology is evolving. Susskind published the book after ChatGPT but before Claude Cowork. It was not until the last 12 months that AI tools have become indispensable to our (university) work. AI platforms such as Cowork are now the medium through which we (or at least I) conduct research, analysis and (often) the creation of internally shared documents and presentations.
What is often misunderstood about knowledge work—and I’d argue the work that higher educational professional staff do—is that the audience for most of our work product is primarily internal. Unlike this book review that you are reading now, where the audience is you (external) and every word is human-written, most of what I create is intended for an internal audience. A deck made in collaboration with Claude is much faster to create than one made slide by slide, and if the purpose of the deck is to move an idea or a project forward and share information internally, then an AI-assisted deck is fit for purpose.
The challenge is that for many professionals (including higher ed professionals), our days are now spent more in conversation with AI than in collaboration with fellow humans. Susskind’s provocation that we should be working through the implications of a future AGI skips over the reality that deep AI collaboration has already arrived.
Speaking for myself, having an AI collaborator is fantastic. Claude never gets tired or annoyed of iterating on a deck, memo, report or spreadsheet over and over and over again. Claude can work on my weird nonstandard workday schedule without complaint. As the accuracy of AI output improves (less hallucination, higher validity and reliability), the utility of the tools advances. I can now create complex financial models and interactive dashboards without assistance from financial analysts or developers, as I can ask the right questions and accurately evaluate the results.
For others, talking with AI all day is isolating, demotivating and exhausting. Tools like Claude Cowork allow us to work much faster and more independently, with the flip side being that everyone is expected to do more work. In our world of online learning and instructional design, I’m particularly worried about what AI will do to the creative and collaborative process of course development and faculty/designer co-creation. We don’t need to worry about a future AGI to be concerned that today’s AI is already making our work different, and in some cases (and for some people) worse.
Reading and discussing How to Think About AI would both catalyze and ground campus conversations on AI and our work. If you are putting together an inventory of books to get your team, unit, division or school to read in order to spur a discussion about AI and your organization’s work (as opposed to AI and teaching and learning), I recommend adding How to Think About AI to your list.
What are you reading?