DISEC - Regulation of Artificial Intelligence: How will we deal with machines that learn?10/28/2022 In the past, technological advancements have been met with a lot of adversity, stemming from a deep rooted fear of change. Whether it was the invention of trains that elicited the fear of the end of pastures, or the creation of the telegraph that people believed to be the work of the devil, the world has always been critical of technological innovation. It just happens to be that now the advancement is artificial intelligence (AI): machines that imitate the behaviour of humans. With this new topic just being introduced, the expression “regulation of AI” has been frequently used. But what does it actually mean? What does it imply? And how will we implement it? History of technological revolution has shown us, that people have always displayed the same behavioural pattern of fear of change: fear that jobs will diminish, that people’s lives will drastically worsen, and that the world will not adapt. The world now displays the same skepticism for the field of artificial intelligence: that all of our jobs will be done by robots, that our privacy won’t exist anymore, that AI will “take over the world”. Redeployment As it has always been, with new technologies being integrated into lives, the more simplistic jobs, like toll collectors, factory workers, or bank tellers, have been replaced by automated machines like ATMs. According to Oxford Economics, since the 2000, 1.7 million manufacturing jobs have been already lost worldwide. In the US alone, robots have taken over 260,000 jobs, which represents 2% of the country’s manufacturing workforce. Furthermore, the McKinsey Global Institute predicts that 375 million jobs will be lost due to automation by 2030. However, even though jobs have been replaced, new, more complex jobs have been created as a result. In addition, since many of the manual jobs will no longer be done by humans, there is more opportunity to work in fields that require social interactions, like teachers. Some of the most likely jobs to increase in demand over the next 15 years include cyber security expert, software developer, data analyst and online content creators. Hence, rather than causing unemployment, AI will create redeployment. Regulations Why should AI be regulated? Technology has been advancing at a much more rapid rate than our ability to react to change. In the past, people have had time to adapt, as the progress of innovation was constant and respective to its time period; but now we have the AI race: the global competition for who will develop the best artificial intelligence first. Because of this contest, and the change it will bring in power dynamics between nations, regulations must be set in place. These rules should focus on the effect of the new technology, rather than the effect of the technology itself. This indicates that instead of trying to regulate the increasingly complex artificial intelligence machines, we need to concentrate on the impact that it will have on our lives: our privacy, our data and everything that these machines will have access to. Although many regulations are already in place in the field of technology, such as data protection, consumer protection and market competition laws, such legislations must be extended to AI as well. One such regulation is regulatory frameworks, necessary to ensure accuracy and a unbiased systems. Even though AI should be an equal system, its predictions are based on the input that is given to them, therefore its outputs will contain the biases and inequalities from the society that they were taken from. For instance, if a company has a white male majority of employees, and uses an AI system to choose new employees based on the current ones, the system will inaccurately predict that the best candidates will all be white males. Furthermore, AI’s require safety critical systems to avoid property damage, financial loss, human injury, human death (etc.), like a self-driving car. There have been already some regulations set in place: the United States’ National Artificial Intelligence Initiative Act of 2020 (became law in January, 2021), which concentrates on development and research of artificial intelligence, the European Union Artificial Intelligence Act (launched in April, 2021), which focuses on systems with an unacceptable risk or high-risk applications, and China’s recent Internet Information Service Algorithmic Recommendation Management Provisions (passed in January, 2022), which concentrates on companies’ use of algorithms. Conclusion Regardless of public skepticism and fear, artificial intelligence will inevitably change our lives as we know them, we just have to be ready for It. Writer: Alice Guzzi Bibliography
0 Comments
Leave a Reply. |
RSS Feed
