REBELMUN 2018 COMMISSION ON SCIENCE AND TECHNOLOGY FOR DEVELOPMENT

Transcription

1 Dear Delegates, As a current undergraduate pursuing a degree in computer science, I am very pleased to co-chair a committee on such a pressing and rapidly emerging topic as this. My name is Jonathon Teague, and in high school, I was both part of the MUN team and the speech and debate team, where I had the pleasure of working alongside my co-chair and intellectual partner in crime, Mackenzie. Our generation is encountering new and unprecedented challenges that raise ethical questions like, When has science gone too far? Are man-made machines going to systematically extinguish mankind? While I do not think that we should halt all research and development into artificial intelligence, I do think that we should be prepared for the inevitable moral and practical consequences to come. I hope we can have a good discussion both on the pros and cons of artificial intelligence and look forward to hearing from all of you this April.

2 Topic Background With leading scientists like Stephen Hawking claiming that artificial intelligence (AI) could spell the end of the human race, it behooves all of us as members of the international community to approach the topic with serious dialogue. In the past decade, rapid advances in machine learning have helped us create self-driving cars and autonomous drones, enabled accurate recognition of speech, translation of text, and natural human-computer interfaces, as well as shown promise in a wide variety of medical applications including drug research and disease diagnosis (Vodrahlli and Bowhmik). The benefits of this emerging technology are diverse and plentiful. However, it is our responsibility as an international committee to ensure that continued applications of artificial intelligence remain ethical and are not used in ways that cause harm, where harm is defined as: physical injury, material damage or actual or potential ill effects or danger ( harm ). Such harm is antithetical to the Sustainable Development Goals this committee seeks to achieve. Modern artificial intelligence development is based upon a concept called machine learning, which is being applied in a variety of interesting ways. Consider the recent and groundbreaking example of AlphaGo. In March 2016, AlphaGo, an artificially intelligent computer system developed by Google s Deepmind group, defeated 18-time world champion Lee Sedol with the score 4-1 in the ancient Chinese game Go (Sang-Hun). What makes this feat so groundbreaking is that AlphaGo was self-taught from scratch by a new form of reinforcement learning where it analyzed thousands of human-played games of Go and simulated games where the artificial intelligence faced itself (Silver). To put it differently, AlphaGo showed the world that machines are now capable of learning and developing algorithms that go beyond human

3 knowledge; AI can teach itself how to do tasks that have never been expressly condoned by human creators and eventually will surpass human intelligence holistically. As machine learning is applied to more sophisticated topics, it is inevitable that the scope of the machine will surpass that of its programmed design. This is further complicated by a host of security issues including hackable vulnerabilities, replacing many facets of human labor and the dual nature of the technology- beneficial applications of AI come with the possibility of maliciously using the same powers for harm (Brundage 16). As AI continues to develop, it is important to protect the technology and the people it affects both from itself and from humans who would seek to use the technology to carry out harm, possibly on a massive scale. In the face of such powerful emerging technologies and the possibility of more sophisticated attacks with the use of AI than the Internet has yet seen, many countries and organizations are looking to AI for the future of cybersecurity. Already, AI is being used as a tool to defend from attacks by organizations and to break into systems via attackers (Lahiri). While the question of what AI is currently capable of doing and what AI will become able to do from here is a serious question to consider, there is a well-known axiom within the realm of cybersecurity: people are much more likely to fail than systems. According to IBM, insider attacks, where an employee or individual who has access to a private network, account for 60% of all cyber attacks and are one of the most difficult to detect. Thus, cybersecurity implicates defense from attacks that originate from the outside and within, a fact which while unchanged by the introduction of AI is now to be made on an unimaginably large scale. An emerging application of artificial intelligence-based technologies is healthcare. Using neural networks developed through machine learning, it has recently become possible to find mutations in genomes and possibly even identify previously overlooked biological phenomena

4 (Maxmen 9-10). This among other medical applications shows that artificial intelligence can be the key to tracking and identifying previously elusive diseases, as well as potentially providing the means to even manufacture cures to terminal illnesses. Further applications of artificial intelligence are still under development, but as the technology continues to play an increasing role in our society, it becomes ever more important to monitor and regulate it.

5 Committee Background The Commission on Science and Technology for Development (CSTD) is a subsidiary of the Economic and Social Council (ECOSOC). Established in 1992, this committee was designed to provide high-level policy advice to the ECOSOC as well as the General Assembly of the United Nations. This commission examines questions related to science and technology and their implications for development. The commission also advances understanding about policy issues concerning emerging technologies and proposes guidelines and recommendations for matters related to science and technology in the United Nations (UNCTAD). Broadly speaking, this committee focuses on areas including investment, international finance, e-commerce, trade, innovation, and emerging technologies (UNCTAD). In particular, the commission places a strong emphasis on the relevance of science and technology in meeting Sustainable Development Goals (UNCTAD). This committee lacks the power to condemn or demand things of specific countries and organizations. Rather, this commission possesses the power to offer recommendations and guidelines as well as policy proposals to the ECOSOC and the General Assembly. Consequently, it is not the responsibility of the CSTD to implement any of its policy proposals or recommendations.

6 UN Involvement While UN dialogue on this pressing issue is still under development, the international community has previously met to discuss the severity of artificial intelligence within the scope of our rapidly changing world. On October 11, 2017, the Second Committee of the General Assembly and ECOSOC met in a joint meeting featuring experts in the field to discuss artificial intelligence in a meeting named The Future of Everything (Chatardová). Under this meeting, general consensus was achieved and summarized by the chair of the General Assembly Second Committee and ECOSOC, Marie Chatardová, regarding the following areas of interest: The future is here. Governments, and by extension the UN, should move from reaction to action, and should devise strategies to manage the risks of technological progress. Risks addressed during the discussions included risks of personal safety, comprised privacy, unwanted consequences, and the need to prevent human bias in technological solutions. Dialogue is crucial. Some participants stressed that the question of how technology can benefit all should not be left to technologists and innovators. Instead, it would require an open and inclusive dialogue among all stakeholders at national and global levels. Technology is a tool, not a vision. The discussion on the impact of technology on society at large needs to be underpinned by a thorough understanding of the problems that technology should solve, and ultimately the vision of how the future we want looks like. People, not machines, come first. Human concerns must be at the centre when designing technology that enables solutions to remain context-specific. Invest in youth. Young people should be equipped with the skillsets needed to embrace

7 technological change, including creativity and innovative thinking. Additionally, lifelong learning will become ever more important, as rapid technological change will further accelerate in an exponential fashion. UN leadership matters. The UN has a key role to play in ensuring that the benefits of new and emerging technologies are distributed in a fair, responsible and transparent manner. The UN should remain engaged in the conversation on the effects of frontier issues on the 2030 Agenda, providing a platform for exchange between all relevant stakeholders.

8 Questions to Consider/ Places to Start: Neuroscientist and philosopher Sam Harris discusses the implications of developing ever more intelligent artificial intelligence: What are some of the benefits unique to developing artificial intelligence? What are their corresponding costs? What should the UN and international community in general do to prepare for the future of artificial intelligence? What can be done under our various political, social and economic situations? Could artificial intelligence help us attain The Sustainable Development Goals ? (See: In what ways could artificial intelligence be used to perpetrate acts of terrorism and potentially on what scales?

9 Works Cited. Brundage, Miles et al. The Malicious Use of Artificial Intelligence: Forecasting, Prevention, and Mitigation. Future of Humanity Institute et al. Feb Accessed 26 Feb Cellan-Jones, Rory. Stephen Hawking warns artificial intelligence could end mankind. British Broadcasting Company. 2 Dec Accessed 26 Feb Chatardová, Marie. Joint Meeting of the Second Committee of the UN General Assembly and the Economic and Social Council. United Nations- Economic and Social Council. 11 Oct Accessed 3 Mar "harm, n." OED Online. Oxford University Press, January Web. Accessed 26 February Lahiri, Kris. How Will Artificial Intelligence And Machine Learning Impact Cyber Security? 15 Feb Accessed 26 Feb Mandate and Institutional Background. United Nations Conference on Trade and Development, UNCTAD, unctad.org/en/pages/cstd/cstd-mandate.aspx. Maxmen, Amy. Deep learning sharpens views of cells and genes. Nature International Journal of Science. Vol 553. Issue Jan Accessed 4 Mar Sang-Hun, Choe. New York Times. 15 Mar Accessed 23 Feb Silver, David et al. DeepMind. 7 April Accessed 23 Feb Vodrahlli, Kailas, and Achintya Bowhmik. 3D computer vision based on machine learning with deep neural networks: A review. Journal of the Society for Information Display. Volume 25. Issue 11. Nov Accessed 23 Feb 2018.