Artificial intelligence, and tools such as ChatGPT in particular, have transformed the way we work. Capable of generating text in seconds, suggesting ideas or answering technical questions, it has become an essential ally for many professionals.
But can we really trust it?
AI: a misleading sense of confidence
We increasingly meet professionals who use ChatGPT to solve complex technical problems. While some suggestions are relevant, others turn out to be completely wrong, leading to costly dead ends in both time and money.
Why?
Artificial intelligence models generate answers based on the information they have learned or can access. They can be extremely relevant, but they can also produce convincing mistakes when they lack context, rely on incomplete information or misinterpret a situation.
Unconscious incompetence: an underestimated danger
In psychology, Martin M. Broadwell’s model is often used to describe knowledge and competence:
- What I know that I know
- What I know that I don’t know
- What I don’t know that I don’t know
The third category is the most dangerous: unconscious incompetence.
It is also one of the main risks associated with AI: producing an answer that appears credible while being based on incomplete reasoning or incorrect information.
And unlike a human being who learns from mistakes, AI suffers no consequences when it gets things wrong. It simply apologizes… before doing it again.
Using AI with discernment
Far from us to suggest that AI is useless.
Quite the opposite.
It is a powerful tool for generating ideas, summarizing documents and automating repetitive tasks. However, it should be used with discernment:
- Always verify the information provided
- Cross-check sources against reliable data
- Do not let AI make critical decisions
- Consult an expert to validate proposed solutions
AI… more human than we think?
In the end, isn’t there something surprisingly human about AI?
That irresistible urge to speak up, to have an opinion on everything… even when there is nothing useful to say.
So the next time ChatGPT offers you a miracle solution, ask yourself:
Is it genuinely relevant, or is it just digital chatter?
What we actually do with AI
At Ingenious Tech, we use AI on a daily basis to:
- prepare documents;
- summarize information;
- support certain analyses;
- draft initial versions of content;
- automate repetitive tasks.
But every output is reviewed, challenged and adapted to the real-world context of the business before it is used.
Copilot, ChatGPT and other AI tools can be extremely useful when used in the right context. Let’s identify what can genuinely bring value to your business.
