The US & the EU on AI Regulation-Sankalpa Koirala (2nd Rank, 1st National Blog Writing Competition)

Humans have created minds greater than themselves in forms of machines. It has helped mankind in almost all the fields today but isn’t it also a threat to ourselves? Isn’t there a possibility of our creations enslaving us? The answer is “yes”. It is a fact the AI has created job displacement and replacement as it has the ability to replace multiple workers. Therefore, regulation in the field of Artificial Intelligence (AI) is of utmost importance today and in the future to come. AI is simply an agent built for a purpose which works through algorithms. AI is an advanced software which deals with data.[1] Such AI can be used to optimize city systems, advance bio-tech, and give data to policymakers. With good policy and governance, AI can be used to make human lives easier but AI always carries a lot of risks which can be broadly divided into risk of misuse (intentional misuse by humans), accident risks (risks arising not from humans and negligence) and structural risks (risks to economy, environment and society). Investment in AI has been increasing and the market was expected to reach $70 billion for AI systems and $83 billion for robotics by 2020.[2] AI lawyers exist today and AIs are able to predict judicial decisions with 79% accuracy.[3] Need for regulation and laws has also been realized by leading AI experts such as Elon Musk.[4] Facial, image and speech recognition, although strengthens law enforcement, it simultaneously raises question about civil liberties, modern wars[5] and possibility of an Orwellian future. Human rights abuses upon the Uyghurs in Xinjiang can be traced to misuse of AIs by the Chinese Communist Party. Facial Recognition gates, cell phone information and government's spyware app in every citizen's phone is common in China. AIs also hold the potential to develop health sectors. Such technology can be used in the present Covid-19 outbreak which can save the lives of many frontline workers but such advancement has still not been accomplished. AIs are now advanced in speech and language,[6] stock market prediction,[7] image recognition,[8] logical understanding[9] and even detecting human emotions,[10] thus calling for regulation in the field. Realizing the potential that AI holds, both in positive and negative aspects, intervention of law is necessary.

AI policy in the US

AI engineering, reliability benchmarks, strength tests, risk management, etc. are important to regulate AI in the modern world.[11] Therefore, it is necessary to determine a standard for AIs in relation to their strength, engineering, abilities, etc. The US and China are in constant competition to make their country a leader in AI technology. Thus, the US has implemented policies so as to maintain their leadership in AI. Their policies are important so as to extract economic gains from the AI. The US congress has been suggested to work on P.L. 101-246, Title IX which restricts exports of crime detection equipment to China, keeping in mind the human rights concern.[12] Even the National Defense Strategy and Department of Defense have considered security of service members and citizens while using their AI for military advancements.[13] They also explicitly mention that AI is not subject to autonomy as it is not a weapon.[14] Therefore, even private entities can be involved in AI development but there are some restrictions.

Policy recommendations such as increasing government funding in AI to $25 billion by 2020, incentivize private sectors with less tax and more data, promotion of STEM education, improvement of collaboration between counterintelligence experts and universities, amendment of labor schedule 'A' occupations list which includes high-skilled AI technologists, creation of visas for AI experts with ten year open market commitments, establish norms for AI have been made to the US congress.[15] There has also been an executive order for maintenance of American leadership in AI that requires provision of access to federal data and compliance with safety, privacy and security.[16] We can see that the US mainly focuses on economic advancement by making the US the central hub of AI and by providing visas. The Obama administration suggested retraining as a solution in case of AI-job displacement, creation of jobs and strong unemployment insurance. Such policies have proved to be weak in the US as seen after the Covid-19 outbreak.

AI Policy in the EU

EU nations have declared plans for co-operation and an expert committee has formed to draft guidelines for ethical use and development of AI. Civil Law Rules on Robotics, as passed by the EU discusses liability in AI-related accidents and misuse, standardization, privacy, data protection and ethics with reference to international human rights. Liability for human act or omission, AI to complement humans and not replace them, 5G connectivity based on the principle of net neutrality, environment friendly use and guidelines for use of medical and care robots has been discussed by the adopted text. Principal's liability is an important concept when it comes to AI. It has also proposed for an agency in the EU which works in the sector of AI. It has recognized Asimov's laws as well.[17] Asimov's Three Laws of Robotics state that (1) robots should not injure human beings, (2) a robot must obey human orders, (3) a robot must protect its existence without violating the first two laws. Additionally, Asimolar principles demand for failure transparency, judicial transparency, value alignment, human control, common good, research, non-subversion, privacy, recursive self-improvement, risk assessment, etc. Furthermore, data localization laws have been passed by many countries which makes data collected in a country to be stored in the country. Such laws come with the price of economic loss.[18] India is debating over data localization while the EU has completely rejected it.[19] Fundamental rights implications of Big Data, as put forward by the EU, discusses the impact on privacy by automated decision making, media freedom and pluralism, justice, training of AIs, cyber bullying and violence, pseudonymization and encryption, fair market competition and data and consumer protection.[20] Data protection helps to protect the privacy of people. Pseudonymization is a procedure through which information of an individual is de-identified for data management and protection. Pseudonymization helps to protect individual data of people. Media pluralism means media should have freedom and the right to have difference in opinions as well. Therefore, the EU aims not for data localization but for its use in a systematic manner in the field of information technology and AI.

Conclusion

Many western countries have been able to develop law and regulate their world of AI. The policies implemented are focused on being the leading country in AI development while still in compliance with International Human Rights. Even India holds a potential to add $1 trillion to the AI market[21] and can significantly contribute in many sectors.[22] India, holding much potential in the sector should strive for maintaining laws like the US. India should maintain itself as a center for AI opportunities and advancement through the introduction of proper laws like the US. With such potential there exists a possibility of abuse and strict law enforcement and policies must be made. Global dialogue in development in AI is important as already realized by the UN.[23] Leading countries in the AI sector must come together to establish ethics and boundaries in the field. Established bodies of the UN have come together to establish regulation in the field of AI with respect to labor laws, environment, human rights, economics, disarmament, economics and development.[24] Uniform laws should be maintained internationally so as to prevent unwanted exploitation. Exchange of ideas in AI can help in its regulation among countries. Therefore, countries should realize that AIs are human creations which should be used for the benefit of mankind and should come up with laws to prevent any unethical use of such great potential. AI invasion to the world of mankind should be limited to movie screens only and for this purpose, uniform laws and policies as discussed, can act like heroes aspiring to prevent greater evil.

[1] Chris V. Nicholson, 'An Introduction to Artificial Intelligence' (Skymind, 2019) [2] Bank of America Merrill Lynch, 'Robot Revolution - Global Robot & AI Primer' (2015) [3] Bex Caygill, 'AI Predicts Outcomes of Human Rights Trials' (2016) [4] Bunz and Janciute, Artificial Intelligence and the Internet of Things: UK Policy Opportunities and Challenges (University of Westminster Press 2018) 14-17 [5] Department of Defense Artificial Intelligence Strategy, 'Harnessing AI to Advance Our Security and Prosperity' [6] Wayne Xiong, 'Achieving Human Parity in Conversational Speech Recognition' (2016) [7] De Spiegeleire, Maas, and Sweijs, 'Artificial Intelligence And The Future Of Defense: Strategic Implications For Small- And Medium-Sized Force Providers' (2017) 43-59 [8] Allison Linn, 'Microsoft Researchers Win ImageNet Computer Vision Challenge' (2015) [9] Cade Metz, 'Google’s Hand-Fed AI Now Gives Answers, Not Just Search Results' (2016) [10] Matt Burgess, 'Microsoft’s AI Can Detect Your Emotions' (2015) [11] Defense of Innovation Board, 'AI Principles: Recommendations on the Ethical Use of Artificial Intelligence by the Department of Defense' [12] Rasser, Lamberth, Riikonen, Guo, Horowitz, & Scharre , The American AI Century: A Blueprint for Action (2019) 26-29. [13] Department of Defense, 'Summary of the 2018 Department of Defense Artificial Intelligence: Harnessing AI to Advance Our Security and Prosperity' 4 [14] Department of Defence, 'Autonomy in Weapon Systems' (2012) [15] Blueprint for Action (n11). [16] Maintaining American Leadership in Artificial Intelligence, Order 13859 2019. [17] Civil Law Rules on Robotics, P8_TA(2017)0051. [18] Matthew Newton and Julia Summers, 'Russian Data Localization Laws: Enriching "Security" & the Economy' (2018) [19] Ronak D. Desai, 'India’s Data Localization Remains a Key Challenge for Foreign Companies' (2019); Protection of Natural Persons with Regard to the Processing of Personal Data and on the Free Movement of Such Data, 119 OJ L (2016); Framework for the Free Flow of Non-Personal Data in the European Union 303 OJ L (2018). [20] Fundamental Rights Implications of Big Data P8_TA(2017)0076; Intelligence and the Internet of Things (n4). [21] Aditi Chandraskar, 'State of Artificial Intelligence in India' (2020) [22] NITI Aayog, 'National Strategy for Artificial Intelligence #AIFORALL', (2018) [23] Audrey Azoulay, 'Towards and Ethics in Artificial Intelligence' [24] 'United Nation Activities on Artificial Intelligence' (2019)