It should not come as a surprise that we at Alfasent believe in the power of AI and how it can support teams in doing their work more efficiently. However, we also believe that it is crucial to understand how AI works to use it more deliberately. Will ChatGPT give you the answer to some basic question you could Google? Sure. Should you use ChatGPT for that? Probably not.
AI has been around for quite some time, but tools like ChatGPT commodified its use. Suddenly it seemed as though AI was everywhere and the world needed to adjust to its presence. And it makes sense. Generative AI tools are powerful. Even with its hallucinations, the first version of ChatGPT was something we had not seen before. Now, three years later, the model has been updated in such a way that basic hallucinations can be avoided with an internet search or “reasoning”. Most of the conversational generative AI tools go beyond just creating text. People can use it for almost any type of task, and that is also what happens. The question remains, should we use it for any type of task?
The verb “to google” has very often been replaced by “to ask ChatGPT”. You will hear people say things like “ChatGPT knows that”. The thing is that ChatGPT does not actually know that. Put very simply, it knows how to create texts that work within the context of the question you asked. Not because it knows the meaning of the words you used, but because it can predict which words are used in that context. The same goes for AI systems that can generate images, videos or sounds. And yes, all of those systems work with an astonishing accuracy.
Risks
One of the risks that come with this accuracy, is that people assume that everything these systems generate is correct. This can lead to rather innocent situations where a certain university lector used incorrect citations in their inaugural speech. But it can also lead to a teenager overdosing because they asked ChatGPT advise about their drug use.
Then there is the risk of unchecked systems, such as Grok, that men love to use to undress women and children, without consent of course, based on pictures of them fully clothed. And while the impact of those generated images is already unimaginable for the victims, there is also the question of what happens with the original picture or other sensitive data that is inputted in those systems?
From using AI to understanding AI
While only one of the above-mentioned examples is actually relevant in the workplace, this does bring us to the point of understanding how AI works. Knowing how these systems work goes beyond understanding how your input results in a specific output. It also entails understanding the risks, for example in terms of data privacy, as well as the impact on the environment.
This is exactly the goal of article 4 of the EU AI Act. To make sure that employees who use AI have a certain degree of “AI literacy”. Or put differently: “skills, knowledge and understanding that allow providers, deployers and affected persons […] to make an informed deployment of AI systems, as well as to gain awareness about the opportunities and risks of AI and possible harm it can cause” (European Commission).
If you know how it works, you will know when to do your own research versus ask ChatGPT a question. You will know when to verify its output. But you will also know which information it needs to give you the best possible output. You will know how to adjust your prompt to reach your goal. You will know what it can and cannot do, so you can adjust your way of working accordingly. You will know how to keep sensitive data safe. You will understand the risks and ethical questions that come with it. And you will have the necessary information to implement an AI strategy that has been thought through and that will actually contribute to your team’s goals.
To help your company reach that goal, we offer a workshop called “Introduction to (generative) AI”. We cover the basics of what AI is and how it works (without getting too technical) but also dive into the risks and ethical aspects. Finally, we look at tools that can be useful for you company and share best practices on how you can use it as efficiently as possible. Always adjusted to the context of your company and your team(s).
Interested to learn more? Feel free to book an introduction call (no strings attached) to see how we can help!