March 29, 2023. Leading tech leaders and scientists are sounding the alarm; the development of AI is moving so fast, we are forced to apply the brakes. The world is not yet ready for an intelligent spitting image. Just when we get used to the concept of AI, the loss of control threatens. It feels as if AI has landed on Earth in a spaceship. And has no intention of leaving. Worrying? We look at it differently. It’s time to seriously study and educate AI.
Our upbringing determines the outcome
AI is programmed to learn as much as it can from us. In this sense, it is like a child who is having experiences, trying to understand what is right and wrong and getting smarter and smarter. It is malleable and dependent on the right input. How we interact with AI now will determine how it will behave toward us in the future. In other words, what output it will give back to us.
In using AI, we are so focused on the output that we forget how important the right input is. The possibilities, as well as the possible negative consequences of AI, depend entirely on our instructions. AI must be trained, and that is only possible when we speak the same language.
Asking the right questions through prompting
So as users, we need to learn AI’s language. In a figurative sense, because we need to understand AI and find a way to communicate both ways. Not only to cooperate effectively, but also to manage the education of AI. In the world of AI, we call this “prompting,” or the way people can talk to AI. Prompt engineering is creating the right questions and instructions to train an AI model and get better results. In doing so, we put responsibility back on ourselves to ask the right question.
Should we teach AI to think like a human?
By learning how to carefully translate our ideas into prompts, we maintain control over the output. We determine what the AI model generates for us in terms of quality, as well as whether the outcome is right or wrong. Through prompting, we teach the system to reason like us. However, the question is: Do we really need to train AI to think like humans? Isn’t this digital brain fundamentally different from ours? This is an essential question as AI models become increasingly intelligent. Speaking the same language is going to help us better interpret and evaluate AI’s responses, but it doesn’t mean you actually understand the model.
AI learns from us, but we must not forget to learn from AI as well. Not just observing, but engaging in the conversation. Learning how to deal with our creation. Only then can we oversee and manage development.
So our question to you as a reader is: are you open to admiring and questioning what AI can do, what you cannot do (well) and see if you can fill in these missing pieces of the puzzle by learning to use AI?
Josefien ten Have – UX Designer, Jasper Waasdorp – Strategist
Related news
You may ask us anything
Or just stop by to say hi. We are thrilled to think about your challenges.
Let us contact you