Is the authorization in the artificial intelligence intended to be collected under a single authority?
International authorities agree that the potential for threats or dangers of artificial intelligence should not be ignored in addition to the many positive potentials. Drawing attention to the extraordinary development of artificial intelligence in the last few years and the results in the areas where it is used, Asst. Prof. Nuri Bingöl says that the development of artificial intelligence technology should be subject to some regulations and controls. Stating that the differences of opinion are about 'how and by whom this control will be implemented', Bingöl emphasizes that this authority should not be considered as the concentration under a single authority.
Üsküdar University Artificial Intelligence Engineering Master's Program Head of Department Assist. Prof. Bingöl made remarks on how rapidly developing artificial intelligence technologies should be audited nationally and globally, what kind of limits should be introduced and what obstacles there are in front of all these.
It is time to use knowledge in a way that maximizes the usefulness of AI
Stating that it is important not to ignore the threat or danger potential of artificial intelligence in addition to the many positive potentials of artificial intelligence, Asst. Prof. Nuri Bingöl said that "Humanity is experienced in this field and should quickly translate this experience into practice. We have learned a great deal from the positive or negative consequences of previous revolutionary technological developments. It is high time to use this know-how in a way that maximizes the usefulness of artificial intelligence.".
There are differences of opinion on how to implement control over artificial intelligence
Stating that many authorities that shape today's societies with their thoughts and even people who play an important role in the development of artificial intelligence share similar thoughts, Bingöl said that "The extraordinary development of artificial intelligence in the last few years and the results in the areas where it is used show that the development in artificial intelligence technology should be subject to some regulations and controls. The differences of opinion are not on this issue, but on 'how to implement this control'.”.
The measures vary according to which stage will be inspected
Referring to the reasons why the debates and differences of opinion are intensified on how to implement the audit, Bingöl explained these reasons as follows: "Artificial intelligence technologies are a great power and there are problems that will come from concentrating this power in the hands of only one institution. These technologies are a two-stage product by nature. The first is to develop the capacity of artificial intelligence, and the second is to use it for a specific purpose. Which of these two stages should be audited or limited? Which stage contains more potential threats or dangers? Naturally, the measures that can be taken according to the answer to this question are changing. The question of 'how to apply it' takes on a different meaning.".
Excessive control and limitation over technology leads to a decrease of its benefits
Drawing attention to the level of control and limitation of artificial intelligence, Bingöl said that "It should not be forgotten that excessive control and limitation on technology leads to a decrease in the benefits arising from the development of that technology. However, there are great existential problems in front of humanity and artificial intelligence technologies have great potential in these matters.".
Bingöl quoted Andrew Ng, one of the important names in the field of artificial intelligence, as saying, "When I think about the existential risks for a large part of humanity: The next pandemic, climate change and the resulting mass population reduction, another asteroid danger... Artificial intelligence will be an important part of our solutions. Thus, if we want humanity to survive and thrive for the next thousand years, let's get AI to work faster, not slower.". Bingöl added that his statements are remarkable.
Audits and limitations should also include the ethical dimension
Stating that the controls and limitations of artificial intelligence should not only have an impact on the threat and danger dimension, Bingöl said that "There is also an 'ethical' dimension as important as these. Even with very small capacity AI products, it can create some personal breaches and threats. In particular, a person, institution or a group of people can be targeted. Therefore, the need for supervision and limitation should not be considered only as parties that produce or use very large capacity artificial intelligence technology. Controls and limitations should be applicable to both large-scale and small-sized AI capacity.".
The steps taken for personal data security and ethics can form a basis for controls and limitations
Stating that it seems to take time for the standardization and regulations of how to implement controls and limitations to be formed, Bingöl said that "While discussions and ideas continue to develop in this field, it is important to put some practical regulations and legislations related to artificial intelligence into past experiences such as the European Union General Data Protection Regulation or the Personal Data Protection Law as applied in our country and to create a managerial framework because the steps taken for personal data security and ethics can form an idea and practical basis on the controls and limitations of artificial intelligence. These legislations, which are built on the source from which data, to what size and for what purpose, how it is protected and shared, are an important element for the capacity building phase of artificial intelligence and can form a control base at least in this regard.".
We should draw the attention of the elements of 'Information Law' to this issue
Stating that another element is the management of information processing power and energy consumption, Bingöl said that "Controls and limitations in this area have been a solution to some of the problems caused by 'Blockchain' technology and have contributed to the technology to be directed to a more positive side. Similar legislation could have an indirect impact on AI technology.".
Pointing out that first of all, fake content produced by artificial intelligence should be detected and what sanctions should be decided, Bingöl stated that "Content that not only contains fake information, but also contains accurate information and is also created by artificial intelligence to manipulate a group or to benefit a group should be prevented. Thanks to the user data collected, all kinds of political and commercial campaigns created by artificial intelligence targeting individuals should be limited so that free will continues to be free. Without wasting time, we should start to establish the relationship between the concepts of 'Access to Data and Use of Information Processing Power' with the existing laws and regulations with an artificial intelligence-oriented perspective. Especially, we should draw the attention of the 'Information Law' elements to this issue.".
It is almost impossible to create a centralized connection and control
Emphasizing the need to supervise a technology that people can easily reach and that can affect their own and others' lives when it arrives, Asst. Prof. Nuri Bingöl said that "However, it does not seem possible to establish a control mechanism on all artificial intelligence models that can be created with only a computer and a little mathematical knowledge.".
Stating that the fact that artificial intelligence technology is subject to an audit and limitation should not be considered as the inevitably gathering of this authority under a single authority, Bingöl completed his remarks as follows:
"In fact, today's technologies focus on decentralized infrastructures and many developments are taking place in this area. Moreover, once artificial intelligence technologies are produced, they can work as embedded systems without being connected to any center or even the Internet, at least for the use phase. For AI systems with this capability, it is almost impossible to create a centralized connection and control, as is seen in every field today.".
Üsküdar News Agency (ÜNA)