Introduction to International Public Policy (IIPP) Haotiain QI SIS-PKU Fall 2019 Week 14 Technological Shock and New Actors December 17, 2019 Part01 C ONTENTS P03 Part02 Part03 Part04 Housekeeping P05 Big Data P27 Artificial Intelligence P61 Q&A Part One Housekeeping Part One Housekeeping 1) Logistics 2) Final Exams Part TWO Big Data Part Two Big Data Part Two Big Data Technology is NOT an apolitical and amoral force. - Many technologists and philosophers have argued that complex technologies are inherently dual-nature, not just dual-use. - This argument goes back to insights developed from decades ago in the field of Science and Technology Studies, explaining that modern technology is essentially co-produced by scientific and societal actors. - Modern technology is not about algorithms as abstract artefacts used for good or bad, but algorithms endowed with a specific structuring function, a certain power and nature. - It is this structuring function, the technological design process itself and the real-world implications that are increasingly under scrutiny within society. - It is also this structuring function that is offering better governance a chance. Part Two Big Data In 2019 alone, global IP traffic reached over 1 Zettabytes, which is equivalent to 1 billion Gigabytes. We are inundated with much more data than we can analyze. Data mining and analyzing has already taken effect: within the government to mitigate fraud, in law enforcement to predict future crimes, in education to improve academic performance, and in infrastructure to monitor the city’s usage of precious resources. Part Two Big Data Widespread use of digital technologies, the Internet and social media means both citizens and governments leave digital traces that can be harvested to generate big data. Policy-making takes place in an increasingly rich data environment, which poses both promises and threats to policymakers. - New digital technologies enable creativity, connect the world, and provide public goods and essential services. - But they also challenge conventional notions of privacy, facilitate crime, and enable surveillance and oppression. How and to which end these data-driven technologies are used is determined by political, corporate, societal, and individual choices. Part Two Big Data Promises - Data offers a chance for policy-making and implementation to be more citizen-focused, taking account of citizens’ needs, preferences and actual experience of public services, as recorded on social media platforms. - As citizens express policy opinions on social networking sites such as twitter, facebook, weibo, wechat; rate or rank services or agencies on government applications, or enter discussions on the burgeoning range of social enterprise and NGO sites, they generate a whole range of data that government agencies might harvest to good use. - Policy-makers also have access to a huge range of data on citizens’ actual behavior, as recorded digitally whenever citizens interact with government administration or undertake some act of civic engagement, such as signing a petition. Part Two Big Data - Data mined from social media or administrative operations in this way also provide a range of new data which can enable government agencies to monitor and improve their own performance, for example through log usage data of their own electronic presence or transactions recorded on internal information systems, which are increasingly interlinked. - And they can use data from social media for self-improvement, by understanding what people are saying about government, and which policies, services or providers are attracting negative opinions and complaints, enabling identification of a failing school, hospital or contractor, for example. Part Two Big Data Examples of good use Global health governance system implemented a flawed prevention system that identified $210.7 million in improper payments to health care providers. Big Data not only helps save money but also saves lives as it expedites the development of new treatments and containment of viral outbreaks. - Earlier in 2018, the World Health Organization declared the Zika Virus a global health emergency and predicts cases to rise to four million in the next year. Currently there are no reliable tests and vaccines for the virus, but utilizing a data driven infrastructure to identify trends and analyze clinical test results will shorten the race to the cure. - Moreover, it can help medical researchers develop new treatments and help doctors make better decisions. For government, or any other non-market players, they can help reap these benefits by solving the data quality and availability issues, to encourage improvement of local electronic system, lowering long-term administrative costs. Part Two Big Data Public safety - Implementing predictive policing is relatively new. In the U.S., it is currently being tested and deployed in 60 cities. The method involves algorithms and data from type, place and time of previous committed crimes in order to assign probabilities of future crime events to regions of space and time. - The city in which the practice was created, Santa Cruz, CA, U.S., saw burglaries drop by 11% and robberies by 27% in the first year the system was implemented. Same success rates were found in Reading, PA, U.S. where crime dropped to the lowest it has seen in 35 years. Part Two Big Data Education - Predictive analysis can be used to identify students at risk of dropping out. Adaptive tutoring system can change according to individual student’s learning styles. Monitoring student retention rates will make way for enhancing student academic performance and therefore overall satisfaction among students, teachers, and the administration. - Data gathered from individual students’ learning styles can also assist teachers as they can adjust their teaching styles. Working from the inside out, this provides insight on which programs are failing or succeeding in which to invest more or less focus and funds on. Part Two Big Data Macroeconomic forecasting In most countries, national statistical agencies collected macroeconomic data through surveys. These surveys are costly and are not available in a real-time fashion. Private sector data sources however have the potential to supplement and even replace traditional surveys with near-real time data. Part Two Big Data Environmental Governance - Versatile: can be used to monitor an area as vast and expansive as the Amazon Rainforest, or it can monitor a small city’s water supply. Enabling environmental protection on every level — individual, community, country and global. - Increased speed and ease of obtaining data. In the past, most environmental data came from individual scientists out in the field. It was a slow, laborious process that didn’t provide useful information for many months or years. With Big Data that same information is gathered much quicker which leads to rapid implementation. Part Two Big Data Deforestation: On the side of companies responsible for deforestation, Big Data provides alternative solutions to the immense tree-cutting done every day — lowering the carbon footprint and decreasing the negative impact on the ecosystem. Endangered species: deforestation also causes problems for numerous plant and animal species. As they lose their natural habitat, the probability that these plants and animals will survive drops significantly. When a forest is razed, a community of plants and animals is also destroyed. By implementing Big Data, entities can effectively monitor both plant and animal species in danger of extinction. Information is gathered and plans are created early enough to fight the problem and preserve the species. Part Two Big Data Poaching: Traditionally, forest rangers and other land protectors have faced an uphill battle in prosecuting poachers because of the enormous areas they’re responsible for. They can only cover small parts of an entire area. With Big Data, however, their reach is significantly expanded. Trouble areas and animals can be pinpointed and actions taken to prevent poaching. Residential waste (smart city): Wasting water is something most cities can’t afford. Cities may be able to monitor how much water you use in a month, and you may be able to monitor when you watered, but other important information goes undiscovered. Big Data is essential in order to forge ahead and provide our progenitors with a sustainable future. The more info gathered, the better protected the environment becomes. Part Two Big Data In short. data has the potential to spur economic growth and improve quality of life in a broad array of fields, by helping individuals and organizations make better decisions. - The private sector is currently using vast quantities of data for variety of purposes, including optimizing energy efficiency in buildings, reducing mechanical failures in equipment, and improving crop yields on farms. - The public sector has many opportunities to use data to address major social issues such like improving health care, fighting crime, building more sustainable communities, and creating more efficient transportation systems. What’re clear: the potential benefits Far from certain: how to achieve those benefits Part Two Big Data Across-jurisdiction Large-scale data analysis is capable of driving global-scale innovation, but data exchange between countries is a major challenge to many potential benefits of data. One major impediment: data residency law and other laws restricting information flow. These laws prohibit data from being stored or accessed outside a nation’s borders. - For instance: Danish and Norwegian data protection authorities prevent the use of cloud services when servers are not located domestically; Canadian provinces of British Columbia and Nova Scotia have instituted laws mandating that personal information in the custody of a public body can only be stored and accessed within the country; In some countries, government data sources are not kept secret. Part Two Big Data Data-driven innovation can be slowed down with these legal restrictions. There is a need for international legal standards for government and private access to data across national borders, initiatives like G8 Open Data Charter, or proposed “Geneva Convention on the Status of Data” - This requires both political and technological cooperation. Policy standards, as well as technological standards - In the area of “Internet of Things”, data formats are being standardized, in most other areas like mobile health, humanitarian aid, the political processes are falling behind. Part Two Big Data Risks Big data presents new moral and ethical dilemmas to policy makers. - For example, it is possible to carry out probabilistic policy-making, where policy is made on the basis of what a small segment of individuals will probably do, rather than what they have done. - Predictive policing can lead to a “feedback loop of injustice”, as policing resources are targeted at increasingly small socio-economic groups. - What responsibility does the state have to devote disproportionately more – or less – resources to the education of those school pupils who are, probabilistically, almost certain to drop out of secondary education? - Consumers trade privacy to allow Facebook to use their data on the basis it will improve their products, but if government tries to use social media to understand citizens and improve its own performance, will it be accused of spying on its citizenry in order to quash potential resistance. Part Two Big Data Efficiency vs rights, a necessary trade-off? Again, surveillance, for example Cameras are everywhere. Children in school, customers in shops, and pedestrians on the street are increasingly under more – and more advanced – surveillance. Today, law enforcement agencies and government entities in many countries make use of facial recognition technology on a routine basis. Private vendors also deploy such systems for commercial use. And demand is ever-increasing in both the commercial and the security sphere: the facial recognition technology market is predicted to be worth $2.67 billion in 2022 in the U.S. alone. But the pervasiveness of these technologies raises many privacy and ethical concerns. Experts and civil liberty advocates criticize the use of facial recognition technologies on citizens and consumers, citing concerns about accuracy, accountability, and the potential for racially disparate impacts of its use. Part Two Big Data what Part Two Big Data Technical AND political solutions For example: research on privacy-preserving data mining. Various techniques can be sued to add noise to sensitive datasets so that individual information cannot be extracted. Like synthetic data, which preserves the usefulness of sensitive datasets by emulating their underlying statistical characteristics while simultaneously masking individual information. -- here comes the issue of organic governance – many agencies are working on this data issue, but their research and management efforts are not coordinated. To achieve a balance between efficiency and rights protection, what’s needed is not just a technological process, but more of a policy framework – a combination of data sharing agreements, incentives for individuals to participate, and robust protection and regulations mechanisms. Part Two Big Data Not just Big Data of Governance But also Governance of Big Data - Voluntary sharing of data is not always assured in private sector. But when there is a strong public interest, public policy should be used to mandate and incentive private sector to share. - In government, the mechanism is different. Data collection is not voluntary. Only limited by the decisions of government agencies and the political process that holds government officials accountable. Some governmental agencies, like those in intelligence community and law enforcement, have a bias in favor of collecting more information. This must be tempered by competing interests, including the civil liberties of individuals and the economic impact of such decisions. Part Three Artificial Intelligence Part Three Artificial Intelligence No other technology has recently generated so many hopes and fears, expectation and trepidation, celebration and condemnation as Artificial Intelligence (AI) In its current form, called deep learning, AI optimizes predictive reasoning by learning how to identify and classify patterns within massive amounts of data. With this super-computing efficiency – being able to simulate myriads of scenarios in seconds – deep learning offers unmatched investigative opportunities, such as comparing genomes within an entire population, recognizing a certain face out what of a crowd, or labelling any location on earth based on millions of pictures. As a response to the development of AI (and big data), the UN Secretary-General António Guterres has just established a High-level Panel on Digital Cooperation to foster a “broader global dialogue on how interdisciplinary and cooperative approaches can help ensure a safe and inclusive digital future for all.” Part Three Artificial Intelligence Looking ahead, future technology seems indistinguishable from magic. But as technology advances, it creates magic. That will be the case with AI. - Arthur C. Clack what Part Three Artificial Intelligence “AI will transform the nature of wealth and power.” Three lenses - The first lens is that of general purpose technology. AI fits the category of general purpose technology, which are classes of technologies that provide a crucial input to many important processes, economic, political, and military, social, and are likely to generate these complementary innovations in other areas. what General purpose technologies are also often used as a concept to explain economic growth, things like the railroad, steam power, electricity, motor vehicle, airplane, or computer. And it’s plausible that artificial intelligence not only is a general purpose technology, but is the quintessential general purpose technology. Kevin Kelly :“Everything that we formally electrified, we will now cognitize. There’s almost nothing we can think of that cannot be made new, different, or interesting by infusing it with some extra IQ.” Part Three Artificial Intelligence - The second lens is to think about AI as an information and communication technology. other technologies in that reference class would include the printing press, the internet, and the telegraph. They make possible new forms of military, new forms of political order, new forms of business enterprise, and so forth. - The third lens is that of intelligence. Unlike every other general purpose technology, which applied to energy, production, or communication or transportation, AI is a new kind of general what purpose technology. It changes the nature of our cognitive processes, it enhances them, it makes them more autonomous, generates new cognitive capabilities. And it’s this lens that makes it seem especially transformative. In part because the key role that humans play in the economy is increasingly as cognitive agents, so we are now building powerful complements to us, but also substitutes to us, and so that gives rise to the concerns about labor displacement and so forth. In addition, innovations in intelligence are hard things to forecast how they will work, and that makes it especially hard to see what it will bring. Part Three Artificial Intelligence what Part Three Artificial Intelligence what Part Three Artificial Intelligence AI and human Substitutive effect vs. Complementary effect - Whether data is ample Whether info is complete Level of certainty Degree of multi-areas what Part Three Artificial Intelligence Labor displacement and inequality. This is not science fiction to talk about the impact of automation and AI on inequality. Economists are now treating this as a very serious hypothesis, and the bulk of belief within the economics community is that AI will at least pose displacement challenges to labor, if not more serious challenges in terms of persistent unemployment. Whether a necessary outcome, debatable. what Part Three Artificial Intelligence Monopoly of tech power Digital services in general, but AI in particular, have a natural global monopoly structure. And this is because the provision of an AI service, like a digital service, often has a very low marginal cost. In a market like that for Netflix or for Google Search or for Amazon e-commerce, the what competition is all in the fixed cost of developing the really good AI “engine” and then whoever develops the best one can then outcompete and capture the whole market. And then the size of the market really only depends on if there’s cultural or consumer heterogeneity. Part Three Artificial Intelligence At the same time, we need to demystify AI: - Still a black box: we simply don’t know Limited resources: big data and computing capabilities, training is difficult Human brain does NOT need these Not a machine version of human intelligence: in some way can be seen as statistical logic plus, based on big data and modern computing power what Part Three Artificial Intelligence Machine Learning tends to work well when: 1. Learning a “simple” function 2. There is lots of data available Machine Learning tends to work poorly when: what 1. Learning complex functions from small amounts of data 2. It is asked to perform on new types of data that it learned from Part Three Artificial Intelligence What machine learning today can and cannot do (the barrier is being overcome) The toy arrived two days late, so I wasn’t able to give it to my niece for her birthday. Can I return it? what “Refund request” Input text Refund/ Support/ Shipping Oh, sorry to hear that. I hope your niece had a good birthday. Yes, we can help with…. Source: Andrew Ng, 2019 Part Three Artificial Intelligence Self-driving car Cannot do Can do what A B hitchhiker stop A 10,000 1. Data 2. Need high accuracy B bike turn left signal 10,000 Source: Andrew Ng, 2019 Part Three Artificial Intelligence what Part Three Artificial Intelligence what Part Three Artificial Intelligence what Part Three Artificial Intelligence what M. Gong, Y. Li, L. et al. SAR Change Detection Based on Intensity and Texture Changes. ISPRS J. Photogramm. Remote Sens., 2014. Y. Li, M. Gong, et al. Change-detection Map Learning using Matching Pursuit. IEEE Trans. Geoscience and Remote Sensing, 2015. Part Three Artificial Intelligence what Part Three Artificial Intelligence what Part Three Artificial Intelligence Ultimate risks We are witnessing a swelling public anxiety about the loss of control to an algorithmic force, which seems to escape our modes of understanding, trust and accountability. Societies are increasingly governed by human-made technologies as much as by the rule of law. One of the principal concerns at the heart of governing AI therefore becomes whether we will eventually find ourselves controlled by powerful technical systems whose design we did not fully understand and whose ramifications we did not anticipate. what In the future, biosensors and algorithms might together capture and analyze an ever more refined record of our biometrics, vital signs, emotions and behaviors. AI will watch, track and evaluate us: we will go from the predictive power of one algorithm to the next. We may unwittingly give up to algorithmic networks unprecedented access to our bodies, genomes and minds and create possibilities for social and bio-control. Part Three Artificial Intelligence Today’s concerns • • • • • • • Privacy Accountability Safety and security Transparency and explainability Non-discrimination Human control Development responsibility Bias and Adversarial Attack what Part Three Artificial Intelligence Adversarial attacks on AI Minor perturbation Hummingbird Hammer what Minor perturbation Hare Desk Source: Andrew Ng, 2019 Andrew Ng Part Three Artificial Intelligence Example: Social bias, facial recognition Joy Buolamwini https://www.nytimes.com/2018/06/21/opinion/facial-analysis-technology-bias.html Tech and system design: embedded biases and discriminations what - Facial recognition - Job interviews - Gender and racial biases - Public safety Part Three Artificial Intelligence “The products of a company called HireVue, which are used by over 600 companies including Nike, Unilever and even Atlanta Public Schools, allow employers to interview job applicants on camera, using A.I. to rate videos of each candidate according to verbal and nonverbal cues. The company’s aim is to reduce bias in hiring…The system’s ratings…reflect the previous preferences of hiring managers. So if more white males with generally homogeneous mannerisms have been hired in the past, it’s possible that algorithms will what be trained to favorably rate predominantly fair-skinned, male candidates while penalizing women and people of color who do not exhibit the same verbal and nonverbal cues.” Part Three Artificial Intelligence “According to the Center on Privacy and Technology at Georgetown Law, the faces of half of all adults in the United States — over 117 million people — are currently in face recognition database networks that can be searched by police departments without warrant. These searches are often reliant on facial recognition technology that hasn’t been tested for accuracy on different groups of people. This matters because misidentification can subject innocent people to police scrutiny or erroneous criminal charges.” “In the case of South Wales, where Big Brother Watch reports that between May 2017 and March 2018 the what faces of over 2,400 misidentified innocent people were stored by the police department without their consent, the department reported a false-positive facial identification rate of 91 percent. But it’s important to remember that even if false-positive match rates improve, unfair use of facial recognition technology cannot be fixed with a software patch. Even accurate facial recognition can be used in disturbing ways. The Baltimore police department used face recognition technology to identify and arrest people who attended the 2015 protests against police misconduct that followed Freddie Gray’s death in Baltimore.” Part Three Artificial Intelligence Leviathan The fundamental problem of politics, is how do you build this leviathan, that doesn’t abuse its power. We have been developing institutions for centuries to discipline the leviathan so that it doesn’t abuse its power, but AI is now providing this dramatically more powerful surveillance tool and then sort of coercion tool, and that could enable leaders of not only totalitarian regimes to really reinforce their control over their country.what It could lead to sort of an authoritarian sliding in countries that are less robustly democratic, and even countries that are democratic. It will also potentially shift power between different groups. Part Three Artificial Intelligence what Part Three Artificial Intelligence Regulation and Governance We’re seeing a lot of work around norm creation and principles of what ethical and safe development of AI might look like. An important example is the European GDPR (General Data Protection Regulation) that regulates how data can be accessed and used and controlled. We’re seeing increasing examples of these kinds of regulations. what Markets do fail and there are profound impacts of new technologies not only on consumer safety, but in fairness and other ethical concerns. Also more profound impacts, like the possibility that AI will increase inequality within countries, between people, between countries, between companies. But regulation can be very problematic. In general, with technology it’s often hard to forecast what the next generation of technology will look like, and it’s even harder to forecast what the implications will be for different industries, for society, for political structures. So designing regulation can often fail. Part Three Artificial Intelligence We need to think about ways to counterbalance the asymmetry of power between the supposed techleaders and the tech-takers. We need different communities to work together. Who are making the efforts? • Transnational and international bodies • National initiatives • Private sector and industries what Part Three Artificial Intelligence Recent examples of Chinese initiatives: • “Six Principles of AI,” Tencent Institute (April 2017) • “Beijing AI Principles,” Beijing Academy of Artificial Intelligence (May 2019) • “Governance Principles for the New Generation of AI,” National Governance Committee for New what Generation Artificial Intelligence (June 2019) • “Artificial Intelligence Industry Code of Conduct,” Artificial Intelligence Industry Alliance (June 2019) Part Three Artificial Intelligence Governance falling behind, but also Embedded: There are people that are conducting AI research, it’s a human endeavor. There are people making decisions, institutions that are involved that rely upon existing power structures. The processes are embedded in policy, and there are political and ethical decisions just in the way that we’re choosing to design and build this technology from the get-go. One of the things that can help is just improving those communication channels between technologists what and policymakers so there isn’t such a wide gulf between these worlds and these conversations that are happening and also bringing in social scientists and others to join in on those conversations. In most cases, though we don’t have concrete policies on the ground yet. But we do have strategies, we have frameworks for addressing these challenges, mapping what’s happening in that space. The hope is to encourage transparency and also collaboration between actors, which we think is important. Part Three Artificial Intelligence New Actors, polycentric AI is really being led by a small handful of companies at the moment in terms of the major advances. Not only the development of technology, but also ethics principles are coming from companies. We will need some external checks on some of the processes that are happening. There certainly is an important role for governments and academia and NGOs to get involved and point out those gaps and help kind of hold them what accountable. All in all, Advances in digital technology are making it possible to collect, store and process ever-expanding amounts of data. This explosion of data and use of data with help of AI technology holds tremendous potential to boost innovation, productivity, efficiency and, ultimately, economic growth and social value. The new technological trend, however, raises many questions: What do individuals think about the data being gathered about their everyday activities (for example, through social media and the internet, sensors, radio-frequency identification chips, geospatial technologies, loyalty cards or transport cards)? Who should own and control such data? What is the right trade-off between privacy, property rights and security and allowing society to benefit from data-driven innovations and better ways of living? Is the right to be forgotten practicable, useful and meaningful, and does it need to be complemented with a right to be remembered? How can society and international community benefit most from big data, artificial intelligence, etc.? Part Three Artificial Intelligence Choose your standards from the list human autonomy harm-benefit ratio fairness explicability robustness and safety explicability agency and oversight privacy and data gov transparency privacy and data governance well-being diversity and non-discrimination transparency responsibility well-being responsibility equitability traceability reliability equitability governability governability accessibility controllability explainability human-centricity what accountability explainability A new paradigm of global governance? Techs and new actors are making governance process more irreducible, more polycentric, more connected… • Emergence: irreducibility • Non-linearity: networked global community • Polycentric networks: states and the rest • Regime complex: material and social authority, complex linkage …… • Regime complex plus? – intersubjectivity and quantum community Part Four Q&A