IT Home On October 28, it was reported that artificial intelligence has taken on the role of assistant in many professional fields. Doctors, researchers, and software developers all use it in their work. The problem is that this leads to what researchers call a decline in skills: people hand over their work to AI, resulting in a drop in their own professional abilities.
IT Home noticed that Kaisa Kärki, a researcher at a research institute in the field of theoretical philosophy, explored this issue in a research paper titled “Who Should Be Responsible for the Decline of Professional Skills?”, which was published in the journal Human Affairs.
“One challenge caused by the degradation of skills is that practitioners have difficulty noticing this change in their daily work. Part of this is due to cognitive biases people have regarding reliance on AI. When we converse with large language models and receive convincing responses, it’s easy to become overly dependent. But these models can also make mistakes.”
Kelki gave an example from the medical industry: If a doctor relies on a large language model but fails to notice the errors presented by the model, it could lead to a medical accident. Once an incident occurs, who should be responsible?
“Generally speaking, philosophy theories related to responsibility suggest that an actor is only responsible for the actions they are aware of. However, the subtle cognitive biases that arise from the use of AI challenge this view: after all, it is difficult to detect our own biases or the tendency to rely too much on AI. Therefore, whether we should be responsible for this silent degradation in capability and the resulting mistakes remains undecided.”
Without AI, can we still function properly?
Kelki pointed out that skill degradation itself is difficult to detect, so it is not appropriate to blame individual practitioners for all the related issues. This is a social and structural problem.
All professional communities need to clarify which are the core competencies of their industry, and these must be firmly in human hands.
Pilots practice the response plan in case of autonomous driving failure repeatedly. Other industries should also consider the same question: If there is no AI when necessary, can professionals complete their work independently?
“In the field of medical education, there are views suggesting that stages should be set for basic skill training where the use of AI is prohibited. For example, diagnostic reasoning is a core ability in medical work, and having a holistic understanding of specialized fields and related operations is also essential.”
Risk needs to be addressed immediately
If AI usage continues to expand in the workplace while the risk of skill degradation is ignored, we may eventually lose the ability to train new people without relying on AI. In other words, people will become deeply dependent on technology and AI, and will lose their ability to recognize its mistakes.
The risk-resistant capacity of human communities is based on a diverse range of human skills.
Kelki said: “The decline in skills poses a huge risk. Before there is a group-wide decline in ability at the industry level, human society and various professional fields should consider this issue in advance.”
Methods for dealing with decline in coping skills
It is necessary to make practitioners aware that using AI may potentially lead to a decline in skills in an subtle manner.
Even experienced and senior practitioners should regularly undergo core competency assessments;
Some work tasks can only be completed by humans, and are not allowed to be handled by AI.
Employers need to establish AI usage guidelines that take into account the risk of capability degradation.
During the higher education stage of professional talents, learning sessions that prohibit the use of AI should be set up;
Design matching incentive mechanisms to encourage people to maintain their professional skills;
In areas where skill degradation threatens critical social businesses, regulations should be introduced to control the use of AI;
Industry communities define the core skills that must be mastered by humans in this field;
In more fields beyond medicine, further research is needed on the cognitive decline issues caused by the use of AI.
Reference material: