Artificial intelligence is profoundly transforming the way we interact with technology. A new paradigm opens in the way of working, creating and developing any type of content and, therefore, the need arises to develop new skills to work with it.

At the same time, it is convenient to review the set of knowledge and skills that humans have used to work until now and identify which of them persist, which expire, and which we should enhance.

From the deterministic to the stochastic (conversational) paradigm

One of the main implications of generative AI is the transformation of the human-computer interaction model. The transition is taking place from a deterministic paradigm, based on the execution of specific commands to obtain unique results, to a stochastic one.

In the latter, based on conversation, the results are not necessarily always identical, although they are equivalent. They are the product of complex statistical calculations carried out by algorithms, impossible to reproduce later.

This puts communication through natural language at the center of interaction with machines, and must be used to outline work strategies and skills to work with artificial intelligence in both academic and professional settings.

Knowing how to formulate clear, complete and unambiguous instructions, critically evaluate the results provided by AI, apply logical reasoning through mastery of language or lead a conversation to obtain the expected results are some of the capabilities that will be achieved. be necessary to enhance in the age of AI.

The risk of losing control

The possibility of creating a text, summarizing an article, generating an image, analyzing a set of data or obtaining a sequence of computer code with a simple verbal request, to name a few examples, carries risks.

One of the most visible is the loss of control by humans, both of the work processes and the quality of the results. This risk is especially accentuated if we delegate the emulation of higher-order thinking skills to AI, thanks to its ability to offer plausible and conventionally acceptable results.

A scenario like this can lead to a progressive abandonment of the performance of these skills by humans, with the consequent detraining and loss of capabilities.

Bloom’s Taxonomy in the Age of AI

However, it does not have to be that way and perhaps this scenario corresponds to a false belief. As we get used to using these types of tools, the hypothesis that the use of artificial intelligence forces people to assume a more expert and active role than one might initially think is gaining strength. Thus, specifically designed skills to work with artificial intelligence would emerge.

In line with this approach, our team has proposed an update for the AI ​​era of Bloom’s well-known taxonomy .

This taxonomy or categorization is a list of objectives (or levels) that evaluate the learning process created in the 1950s by Benjamin Bloom , psychologist and pedagogue at the University of Chicago. His hierarchy orders thinking skills into six categories starting with lower order skills – such as remembering or understanding – to those the author considers higher order – such as evaluating and creating.

Skills to work with artificial intelligence

In this new version, actions and tasks that humans are beginning to perform with the help of AI have been incorporated, such as designing complex instructions ( prompting design ), integrating unexpected discoveries and results or identifying and counteracting biases, all of this. operating with natural language.

Although for the moment it remains a theoretical proposal, the work of these specialists highlights the need to continue applying all thinking skills, especially those of a higher order, to carry out any of the identified actions.

If we analyze it carefully, we will realize that, even in actions that were traditionally considered lower level, such as remembering or understanding, with AI these also require an expert approach. Otherwise it is not possible to guarantee quality results adjusted to the desired objectives.

AI as an assistant

Bridging the gap between the human and the artificial, to refer to the interaction with a generative AI interface we can use the metaphor of the assistant. Working with an assistant carries great responsibility since it is we, and not our assistants, who are ultimately responsible for the quality of the work.

To do this, it is necessary to apply skills to work with artificial intelligence that are creative to conceive or design the final product, plan the work process, evaluate the quality of intermediate results, make decisions during the process or transmit new directions to our collaborators to that they can carry out the assigned tasks according to our expectations. If necessary, we will also share background information or our own knowledge with them.

Critical capacity, analysis and evaluation

To carry out this complex task, it is not only necessary to have a high level of expert knowledge, but also, among the skills to work with artificial intelligence, we must constantly apply critical thinking. Preserving critical capacity in a context where AI can offer plausible results, but not necessarily based on a human thought process, is essential to evaluate and filter the information provided by the machine.

This ability is located in the high ranges of Bloom’s taxonomy. It is linked to analysis and evaluation skills and is essential to be able to make informed decisions and avoid possible biases.

In this sense, students and professionals must learn to critically dialogue with AI, assuming a leadership role in the creation process and taking responsibility for the results obtained.

Author Bio: Xavier Mas García is a Specialist in Digital Education at UOC – Universitat Oberta de Catalunya