Anyone who has written a thesis or undertaken serious research knows that it feels like an endless endeavor. The more information you find, the more questions you have, and the stronger the feeling grows that you still know nothing. The metacognitive awareness we develop leads us to realize how much we still have to learn.
The opposite is also true; the less we know, the more we think we know. People with low skill or knowledge in an area tend to overestimate their competence.
None of what I’m saying is new. It’s a cognitive bias identified by psychologists David Dunning and Justin Kruger in 1999, called the Dunning -Kruger effect. It’s observed in many areas and demonstrates how a lack of self-criticism and metacognitive awareness can distort our perception of reality.
In the context of business (and especially in innovation), the effects of this bias can be disastrous.
If our knowledge of a subject is limited, we may overestimate our competence and our ability to make sound decisions. We dive headfirst into a new product, service, market, or technology believing we know enough, when in reality, we are taking a huge risk.
What prevents us from making more informed decisions?
Until recently, the limit was the human capacity to process enough data to truly understand what’s needed to make sound decisions. Because so much information is available, and humans have a limit to their processing capacity, reaching that point usually takes time. Enter artificial intelligence. What used to take months now takes only a few hours. The entire human library, all the accumulated knowledge of thousands of years, is just a few clicks away. And it doesn’t just allow us to learn something more easily and quickly; it even leads us to question the need to know it in the first place. Why do we need to be experts in a subject if AI is already competent and works for us? For example, why do we need to know about web design if AI can design a website for me in minutes?
A highly dangerous cocktail is being created: the ignorant empowered by AI. Even before, due to the Dunning -Kruger effect, this person felt they knew everything. Now, they think so even more, since they have AI. They now know about web design, not only because they believe they know when they actually don’t (they’re incapable of judging whether something is well done or not), but because they are now capable of doing it. They are capable of creating a website (in a matter of minutes, even). And since they manage to do it, they’ve acquired the skill, right?
No. The truth is, there’s a huge gap between the output generated by an expert using AI for a task they know well, and the output generated by an AI operating independently, controlled by someone who has no idea what they’re doing. In the first case, a human-machine validation exchange takes place, where the human supervises and provides feedback on the AI’s workflows. This improves the overall result, creating a collaboration between human and AI capabilities. This has been called Human-In- The-Loop (HITL), and has demonstrated enormous impact on results, especially in high-risk sectors (health, finance), complex or ambiguous tasks (innovation, strategy) and early-stage or rapidly evolving systems (technology).
That’s why:
- It’s still important to seek an expert if you want something done right, or
- It remains valuable to study a subject and acquire competence, even if it takes time. At the very least, its important to gain the knowledge of what works and what doesn’t, and what to consider.
- It is necessary to learn AI, to understand and take advantage of its benefits, but also its limitations.
Blaster In- Sight
At Blaster, we believe in the power of combining human expertise with technology, be it AI, IoT , Big Data, or anything else. We developed Blaster In- Sight , an investigative service that combines the accumulated experience of our research team (over 14 years) with the power of software we created in partnership with d_eye , based on analytics and artificial intelligence. This allows us to shorten timelines and reduce costs, increase data processing capacity, and maintain that crucial validation (Human-In-The-Loop) which gives us the confidence to know that we are delivering a top-quality result. Write to us if you want to know more.
