What Happens When Everybody Has Access to Knowledge?

An interesting article by Azeem Azhar on Bloomberg, “AI will upend some basic assumptions about how companies are organized”, reflects many of the points I made in an article a couple of days ago about the future of consulting and how it will extend to just about every other business sector.

The role of LLMs created by generative artificial intelligence has been, fundamentally, to facilitate access to knowledge, in much the same way that search engines did two decades ago.

When search engines began to work properly, they stopped using simplistic and easily manipulated criteria such as word frequency, opting instead for mechanisms based on attention as an indicator of relevance (actually adapting the citation metrics that Larry Page and Sergey Brin knew well from their time as doctoral students. This simplified the task of accessing knowledge, and we went from having to go to a library and rummage through shelves to being able to access knowledge from our computers.

For those of us who started our doctorates in the mid-90s, the change was impressive because we we lived through it: we started out Phd by going to the library to do our first review papers, but we ended by writing our doctoral theses pretty much from our desks. As the internet spread, publishers began to make their papers accessible online for those with a license, while newspapers, magazines, books, etc. were digitized; overnight, access to knowledge changed completely, requiring a different kind of effort and skillset.

This new kind of access does not necessarily mean things got easier. When all knowledge became just a mouse click away, the skills needed for research also changed: nobody believe that anyone with an internet connection could write a doctoral thesis. It simply didn’t work that way: the tool was important… but you had to know how to use it.

It’s a similar story with LLM and AGI. OpenAI’s Deep Research, Google’s AI co-scientist and others don’t just answer a prompt with a complete, reasonably rigorous response and fewer hallucinations, they also include a complete report explaining how they reached their conclusios, and can also purposes a bibliography. As a result, it is tempting to think that anyone can now be a researcher or a consultant, given that access to information is pretty much universal.

However, once again, this is not the case. Much of the world we know is based on the idea that access to knowledge is scarce and expensive, and that accumulated experience is also necessary to be able to carry out certain value-added work. Now, AI is making access to knowledge even more abundant than search engines did, generating answers that require less effort to understand, and that could make it seem as if no experience were needed to use it.

If you work in education at levels where your students already routinely resort to AGI and where it makes no sense to forbid using it, as is my case, it seems obvious in light of the experience of the last couple of years that access to universal knowledge and to certain sophisticated tools that organize it for you is by no means a guarantee that the student is going to do a good job.

AGI has gone from stochastic parrot to deep thinker in a very short space of time, but that doesn’t mean that anyone, with these tools, can become a consultant, teacher, engineer or researcher. Believing so is potentially dangerous. These tools can certainly work wonders, but to get the most from them requires practice and experience. In other words, they cannot replace a good education and the right kind of experience.

Whatever kind of organization you run, if you put sophisticated tools in the hands of people without the right experience or knowledge, it is very likely that many of the decisions they make will be incorrect. It is not the tools, it is knowing what to do with them, how to organize their results — beyond copying and pasting them — and, above all, knowing what to ask and how to ask it.

In conclusion, as knowledge becomes practically free, the question is no longer about how to access it, but knowing what to do with it; it’s about organizations and individuals understanding how to ask the right questions, evaluating the answers and acting accordingly. And contrary to what many people think, all this is going to be harder than it might at first seem.

(En español, aquí)

And for those of you who might have found this article interesting, here’s an additional 12-minute podcast highlighting its most important elements, in a conversational format:

What happens when everybody has access to knowledge⁠? by Enrique Dans for dummies

Enrique Dans’ article explores the impact of readily available knowledge, facilitated by tools like search engines and…

creators.spotify.com

—

This post was previously published on MEDIUM.COM.