Threats and challenges of AI
The growing dependence on AI systems also presents potential risks.
UN IA UND
The under-utilization of AI is considered a major threat: missed opportunities for the EU could mean a poor implementation of major programs, such as Green Agreement, losing a competitive advantage towards other parts of world, economic stagnation and the poorer possibilities for people. Sub-utilization could result from public distrust and companies in AI, mediocre infrastructure, lack of initiative, low investments or, since automatic learning of AI depends on data, fragmented digital markets .
Surutization can also be problematic: Investing in AI applications which prove not to be useful or applying AI to tasks for which it is not suitable, for example by using it to explain complex societal problems .
Responsibility: Who is responsible for the damage caused by AI?
An important challenge is to determine who is responsible for damage caused by a device or service operated by AI: in an accident involving an autonomous car. Should damage be covered by the owner, the automaker or the programmer?
If the producer was absolutely free from responsibility, there can be no incentive to provide a good product or good service and that could affect people’s confidence in technology; But the regulations could also be too strict and stifle innovation.
Threats of AI to fundamental rights and democracy
The results that AI produced depend on how it is designed and the data it uses. Design and data can be intentionally or involuntarily biased. For example, certain important aspects of a problem may not be programmed in the algorithm or can be programmed to reflect and reproduce structural biases. In addition, the use of numbers to represent complex social reality could ensure that AI seems factual and precise when it is not the case. This is sometimes called mathematical wash.
If it is not done correctly, AI could lead to decisions influenced by data on ethnicity, sex, age when hiring or shooting, loans or even in Criminal procedures.
AI could considerably affect the right to confidentiality and data protection. For example, it can be used in facial recognition equipment or for online monitoring and profiling of individuals. In addition, AI makes it possible to merge information that a person has given on new data, which can lead to results that the person would not expect.
He can also present a threat to democracy; The AI has already been blamed for having created online echo rooms according to a person’s previous online behavior, only displaying the content that a person would like, instead of creating a public debate pluralist, just as accessible and inclusive. It can even be used to create a false video, sound and extremely realistic images, called Deepfakes, which can present financial risks, harm the reputation and a decision -making decision. All this could lead to separation and polarization in the public sphere and manipulate the elections.
The AI could also play a role in the prejudice of freedom of assembly and protest, because it could follow and profiled individuals linked to certain beliefs or actions.
Impact of AI on jobs
The use of AI at the workplace should result in the elimination of a large number of jobs. Although AI should also create and make better jobs, education and training will have a crucial role in prevention of long-term unemployment and ensure qualified workforce.