Some time ago I saw the following cartoon: On the TV news comes the news that three people have died at home because the room temperature was 70°C and all the windows and doors were locked. Next to the shocked viewer of the news was a small voice assistant, such as those now available from many major providers, which uttered a quiet "Hi hi hi!".
I think this cartoon captures very well the potential threat that could be hidden in the more and more widespread aids with artificial intelligence. And in the future, artificial intelligence and assistance systems should relieve us even more in all situations of life. At the same time, we must realize that we are opening further doors to monitor, control and manipulate our everyday lives. In addition, these little helpers also offer new attack vectors for unfriendly contemporaries, whether artificial or flesh and blood.
An already older study* on the potentially malicious use of artificial intelligence paints the real threat potential very drastically. The study describes possible scenarios and how they could be countered. The study concludes that we are challenged at all levels to quickly address the consequences of AI technology: Researchers and developers must not just focus on the desired functions, but must think very carefully about how to avert system errors, operating errors and targeted attacks.
Secure AI development with Ginzinger
Legislators must quickly seek advice from AI experts so that they do not fall decades behind with the necessary regulations. And ultimately, each individual user is called upon to consider the extent to which he or she wants to be helped by AI systems. As a technician, the new possibilities are of course highly appealing. However, right from the start, you should weigh the true benefits of a new intelligent helper against the dangers, acquire competence in dealing with AI systems, question how the thing works and not allow yourself to be artificially incapacitated. If in doubt, it's better to switch it off!