If Generative Algorithms Are Going to Work for Us, We’re Going to Have to Learn How to Use Them

Until very recently, few people knew about generative algorithms, but they are rapidly becoming part of the new working reality for growing numbers of people, in much the same way as word processors and spreadsheets.

OpenAI’s release of ChatGPT Enterprise, which follows the addition of generative tools of the same type to Microsoft’s suite, promises to revolutionize the skills required for many jobs. The fact that there are no fears about what a tool of this type does with corporate information, incorporates centralized administration features and can be used within companies with a certain degree of confidence and security, will lead to more and more organizations demanding that candidates are fluent use of generative algorithms and that existing workforces get up to speed very quickly.

As I saw during last year’s classes, using generative algorithms is not just about opening ChatGPT and asking a question. That’s not a skill. Using it properly involves a lot of details ranging from how to ask questions (prompt engineering, etc.) to how to use it or how to make sure about the reliability of its answers, among many other things. But above all, it implies that companies and workers understand what should and should not be asked of this type of tool: given that the lack of use of a tool often leads to people losing their skill, we must reflect on what we want our employees to do with generative algorithms and what we done in the traditional way.

We should not forget that, although it may seem to my readers that generative algorithms have been with us for a long time, the reality is that few people as a whole use them or have even a basic understanding of what they are: the generative algorithm revolution is with us, but as is often the case with new technologies, widespread adoption is slower than many think.

The battle around algorithm training is proving to be extremely interesting. The media has to decide whether to cede their huge repositories of information to technology companies to train their algorithms, or whether they prefer to try to create these tools themselves so as to offer advanced functions to their own users. This process goes way beyond the shortsighted copyright, and evolves more towards specialization.

For Stephen King to consider whether or not an algorithm trained with his books (which is actually a bit scary 🙂 could exercise creativity and come up with a whole that is better than the sum of its parts is an interesting mental exercise: those of us with data archives (in my case, 20 years of documents on technology and its impact in two different languages) might consider the possibilities of an algorithm trained with our material.

We need to understand that these tools are by no means limited to copying and pasting, but can carry out statistical processes such as inference, correlation or Bayesian statistics, very similar to those carried out by our brains when learning. Belittling the algorithm and reducing it to a supposed role of “stochastic monkey” may be a gross mistake.

Could I ask my algorithm to write my daily article or my columns for various media? Possibly, but I don’t see myself doing so (moreover, I would see it as fraud), much less do I see myself asking it to choose, from the news of the day or week, which of them to write about. Instead, it seems perfectly reasonable to me to ask it to proofread a finished article, to check for typos or grammar, or even to research certain data and provide me with sources to document it.

In reality, the choice is clear, and has been repeated many times throughout history: we accepted that children should use calculators, but we did not want them to stop learning basic mathematical operations. We use search engines, but there is a long way between using them to find something and outsourcing critical thinking and simply accepting the first result they show us. We should be clear about what we want these kinds of tools to do, and what we think it is better to keep doing ourselves.

The increasing availability of this types of tools promises interesting times and challenges ahead. We must avoid the urge to simply let them take over, and instead give serious thought to how to balance sustainability and convenience, so that these tools end up working for us instead of us working for these tools.

(En español, aquí)

—

This post was previously published on MEDIUM.COM.

***

You Might Also Like These From The Good Men Project

Compliments Men Want to Hear More Often

Relationships Aren’t Easy, But They’re Worth It

The One Thing Men Want More Than Sex

..A Man’s Kiss Tells You Everything

—

Photo credit: iStock.com