A.I.’s un-learning problem: Researchers say it’s virtually impossible to make an A.I. model ‘forget’ the things it learns from private user data

assassin_aragorn@lemmy.world to Technology@lemmy.world – 591 points – A.I.’s un-learning problem: Researchers say it’s virtually impossible to make an A.I. model ‘forget’ the things it learns from private user data
finance.yahoo.com

I'm rather curious to see how the EU's privacy laws are going to handle this.

(Original article is from Fortune, but Yahoo Finance doesn't have a paywall)

210

You are viewing a single comment
View all commentsShow the parent comment

I'm just going to leave this here.

some random article

A quote from the article, I found especially interesting.

"As a result, no one on Earth fully understands the inner workings of LLMs. Researchers are working to gain a better understanding, but this is a slow process that will take years—perhaps decades—to complete."

Quite an interesting read and I'm sure you can find some others if you want to and try hard enough.

This is a somewhat sensationalist and frankly uninteresting way to describe neural networks. Obviously it would take years of analysis to understand the weights of each individual node and what they're accomplishing (if it is even understandable in a way that would make sense to people without very advanced math degrees). But that doesn't mean we don't understand the model or what it does. We can and we do.

You have misunderstood this article if what you took from it is this:

It’s also very similar in the way that nobody actually can tell precisely how it works, for some reason it just does.

We do understand how it works -- as an overall system. Inspecting the individual nodes is as irrelevant to understanding an LLM as cataloguing trees in a forest tells you the name of the city to which the forest is adjacent.