WHEN CODE IS LAW, ALGORITHMS MUST BE MADE TRANSPARENT

by

Dirk Helbing (ETH Zurich/TU Delft/Complexity Science Hub Vienna) and Peter Seele (USI Lugano)

 

The rule of law as one of the key pillars of open democratic societies is currently challenged by private companies digitally shaping societies. In times where algorithms determine more and more what can and cannot be done in everyday life, “code is law” (1). Currently, however, code is neither being passed by parliaments, nor do the people have a say in how the algorithm-driven world is going to work.

As put forward in 1998 under the headline of “lex informatica” (2), technology itself has become a regulator. Following the original argument, it is code that determines the freedoms of the individual as well as of the legal system: “The importance of our commitment to fundamental values, through a self-consciously enacted constitution, will fade. We will miss the threat that this age presents to the liberties and values that we have inherited. The law of cyberspace will be how cyberspace codes it, but we will have lost our role in setting that law” (3). 18 years later, we see that this has become true on the scale of entire societies. In particular, in today’s attention economy, nudging, neuro-marketing, scoring, social bots, personalized pricing and AI-based content filtering have undermined open democratic discourse, i.e. the very basis of deliberate democracies built on collective intelligence, participation and openness.

Besides chat bots and personalized messages steering public opinion and manipulating elections with the help of social media (see the case of Cambridge Analytica and Facebook), it is a problem of today’s surveillance capitalism (4) that a few large Internet companies download a very detailed picture of our lives for free and give us little economic benefits and little choice of how this data is being used. This turns people into objects, which contradicts human dignity, and gives rise to obscure business models and misuse, which discriminate people and disrespect human rights.

“Creative destruction” as postulated by J. Schumpeter (5) as one of the key pillars of capitalism seems fine, but it must happen within reasonable limits. After the experience of World War II, the Third Reich, and the Holocaust, we cannot allow it to shake the very foundations of civilized life. But as law makers struggle to keep up with the pace of the digital revolution and its disruptive changes, how can we make sure code will be working in the best interest of humanity and all of us?

Restrictive regulations would slow down innovation. A similar thing would apply, if a new authority would have to approve algorithms before their deployment, if they may interfere with the way society evolves. The only way to manage the challenges of the digital age at the pace of digital innovation is algorithmic transparency. But how to achieve it without distorting competition in a free market society subscribing to deliberative democracy? Based on promises and self-declarations of companies? Certainly not. Just recently it has been proven that big tech companies will not change, until government steps in (6). Self-regulation as proposed by private actors in industry is increasingly critiqued as ineffective, even as “having burglars install your locks”, as put forward by Rob Moodie et al. in an interview (7) based on a Lancet study on the influence of company lobbying on public goods (8). 

Some progress is already on the way. Non-governmental organizations (NGOs) like AlgorithmWatch are concerned about algorithmic decision making (ADM), particularly its inherent dangers. AlgorithmWatch calls algorithmic decision making procedures a “black box” and, therefore, they have put together “the ADM Manifesto”, stating that the “creator of ADM is responsible for its results. [But] ADM is created not only by its designers” (9). The debate about creation and responsibility reveals the challenges in times when some algorithms already create other algorithms, while the question of responsibility and liability requires the existence of a legal entity. 

Given the legal, ethical and commercial difficulties in governing algorithms, we plea for algorithm transparency with a delay, based on the legal construct of intellectual property right protection. In analogy to patent protection, we propose algorithmic protection – but given the speed of digital innovation for a period of at most 24 months rather than decades. Within this time period, companies would typically make 95 percent of their profits, and new software versions would come out. After 24 months, the code would be unlocked and made open source. It is suggested, however, that exceptions apply, for example for code that touches national or cyber security, which would need separate quality and security control mechanisms. For all other code we suggest that companies, scientific institutions, NGOs and/or civil society would check whether the algorithms were consistent with human rights and with the values of our societies, or whether they discriminated, manipulated, obstructed, or harmed people. In such a way, violations of data protection laws, the discrimination of people (e.g. by certain personalized pricing schemes), or breaches of human rights would be revealed, such that feedback loops would set in, promoting better quality standards in the future. This would support a design for values (10), as they are laid out in our constitution, the Universal Declaration of Human Rights, or the UN Sustainable Development Goals. The IEEE, the biggest organization of engineers worldwide, supports a similar approach by demanding ethically aligned design (11). 

As a further benefit of algorithmic transparency with a delay, everyone could learn from each other’s code. This would promote combinatorial innovation, which could benefit everyone and may lead even to the prevention of conflict and the promotion of peace (12). It would also be the basis of a true information and innovation ecosystem, in particular if all personal data would be made accessible based on the principle of informational self-determination (13).

Many billionaires have recently decided to donate half of their fortune. It is time to extend this philanthropic principle to algorithms and data. Small and medium-size businesses, spin-offs, scientific institutions, NGOs and civil society can only make significant contributions to a better future, if they get access to sizable amounts of data and powerful ways of processing them.[1]

In accordance with one of the UN sustainable development goals, following our proposal would lead to an inclusive digitization, the “digitalization 2.0” (14). Given the serious sustainability crisis of our planet, which threatens one sixths of all species (15), it is our responsibility to unlock the potentials of data and algorithms to the benefit of our planet and the species living on it. In times, where the Earth is geared towards global emergencies, which puts many lives at risk, we must promote more resilient forms of society and more cooperative forms of innovation. Opening up algorithms after 24 months and establishing full informational self-determination when it comes to our data (13) is a feasible approach, which can largely accelerate the progress of humanity towards solving its existential problems and achieving a higher quality of live for everyone. What keeps us from doing this now?

1. 1. Lessig, L. Code and Other Laws in Cyberspace (Basic Books, New-York, 2000).
   2. Reidenberg, J. R. Lex informatica: the formulation of information policy rules through technology. Texas law Review 76, 553-594 (1998).
   3. Lessig, L. Code is law: on liberty in cyberspace. Harward Law https://harvardmagazine.com/2000/01/code-is-law-html (2000). 
   4. Zuboff, S. A digital declaration. Frankfurter Allgemeine http://www.faz.net/aktuell/feuilleton/debatten/the-digital-debate/shoshan-zuboff-on-big-data-as-surveillance-capitalism-13152525.html (2014). 
   5. Schumpeter, J. Capitalism, Socialism and Democracy. Routledge, London, (1994) [1942]. 
   6. Mahdawi, A. Google’s snooping proves big tech will not change – unless governments step in. The Guardian https://www.theguardian.com/commentisfree/2018/aug/14/googles-snooping-proves-big-tech-will-not-change-unless-governments-step-in (2018). 
   7. Oswald, K. Industry involvement in public health “like having burglars fit your locks”. MedwireNews https://www.news-medical.net/news/20130215/Industry-involvement-in-public-health-e28098like-having-burglars-fit-your-lockse28099.aspx (2013). 
   8. Moodie, R et al. Profits and pandemics: prevention of harmful effects of tobacco, alcohol, and ultra-processed food and drink industries. The Lancet 381, 670-679 (2013).
   9. Algorithm Watch https://algorithmwatch.org/en/the-adm-manifesto/ 
   10. Design for Values, http://designforvalues.tudelft.nl/ 
   11. IEEE Global Initiative. Ethically aligned design Version 1 and 2. http://standards.ieee.org/develop/indconn/ec/ead_v1.pdf (2016) and http://standards.ieee.org/develop/indconn/ec/ead_v2.pdf (2018). 
   12. Helbing, D., Seele, P. Sustainable development: turn war rooms into peace rooms. Nature 549, 458 (2017). doi:10.1038/549458c 
   13. Helbing, D. How to stop surveillance capitalism. The Globalist https://www.theglobalist.com/capitalism-democracy-technology-surveillance-privacy (2018). 
   14. Helbing, D. (ed.) Towards Digital Enlightenment (Springer International Publishing, 2018). 
   15. Urban, M. C. Accelerating extinction risk from climate change. Science, 348, 571–573 (2015).

 

---

[1] To avoid misuse, however, access to data, code, and functionality should be proportional to qualification and a reputation for responsible use.

IS THE “MORAL MACHINE” A TROJAN HORSE?

by Jan Nagler 1, 2 and Dirk Helbing 2,3,4

How should self-driving vehicles faced with ethical dilemmas decide? 

This question is shaking the very foundations of human rights.

In the “Moral Machine Experiment” (1), Awad et al. perform an international opinion poll on autonomous vehicles. While the authors emphasize not to blindly follow local or majority preferences, they highlight challenges that policymakers must be aware of if special groups of people are not given a special status. This may push politicians to follow popular votes, while car manufacturers already pay attention to opinion polls (2).

However, is a crowd-sourced ethics approach appropriate to decide, whether to prioritize children over elderly people, women over men, or athletes over overweight persons? Certainly not. The proposal overhauls the equality principle, on which many constitutions and the UN Charter of Human Rights are based.

While we acknowledge that laws have to be adapted and upgraded to account for emerging technologies, and that moral choices may be context-dependent, changing the most fundamental ethical principles underlying human dignity and human rights in order to more successfully market new technologies may result in a rapid erosion of the very basis of our societies.

Giving up the equality principle (as Citizen Scores do) could easily promote a new, digitally based feudalism. Moreover, in an unsustainable, “overpopulated” world, “moral machines” would be Trojan Horses: they would threaten more human lives than they would save. Autonomous AI systems (not necessarily cars or robots) would potentially introduce principles of hybrid warfare to our societies.

Instead of just managing moral dilemmas, we must undertake all reasonable efforts to reduce them. Therefore, we propose that autonomous and AI-based systems should conform with the principle of fairness, which suggests to randomize decisions, giving everyone the same weight. Any deviation from impartiality would imply advantages for a select group of people, which would undermine incentives to minimize risks for everyone.


(1) E. Awad et al., The Moral Machine experiment, Nature 562, 59-64 (2018)
(2) A. Maxmen, Self-driving car dilemmas reveal that moral choices are not universal, Nature 562, 469-470 (2018)


Affiliations:

1. Frankfurt School of Finance and Management, Adickesallee 32-34, Frankfurt, Germany
2. Computational Social Science, Department of Humanities, Social and Political Sciences, ETH Zurich, Clausiusstrasse 50, CH-8092 Zurich, Switzerland
3. TU Delft, Faculty of Technology, Policy, and Management, The Netherlands

4. Complexity Science Hub, Vienna, Austria

E-mail addresses: j.nagler@fs.de; dhelbing@ethz.ch

Comment on Awad et al., The Moral Machine Experiment, Nature 562, 59-64 (2018);Link: https://www.nature.com/articles/s41586-018-0637-6

Open Source Urbanism: Beyond Smart Cities

Sergei Zhilin (TU Delft), Jeroen van den Hoven, Dirk Helbing (ETH Zurich/TU Delft/Complexity Science Hub Vienna)

Open Source Urbanism can help mitigate the migration crisis and improve living conditions all over the world.

The dream of building “good cities” is old¹. Since the 20^th century, there have been many attempts to create, develop or shape cities, sometimes even from scratch. Examples range from gigantic modernistic approaches known from Brasilia and Chandigarh, to more radical, but theoretical concepts aimed at changing society and engineering social order, such as Ecotopia or the Venus project. Recent developments are driven by the planetary trend towards urbanization, mass migration, and the need for sustainability. New visions of a global urban future were developed, such as “Sustainable”, “Eco”, or “Resilient” Cities, typically based on a top-down approach to the design of urban habitats.

Cities created from scratch heavily depend on massive private investments, for example, Songdo in South Korea or Lavasa in India. Despite ambitious goals and many technological innovations, their long-term success cannot be taken for granted, as they are often conceived by urban planners without the participation of people who later live in these cities. Such projects are typically implemented without much feedback from citizens. This makes it difficult to meet their needs. In fact, some of these cities have ended as “ghost cities”.

In the wake of the digital revolution, data-driven approaches promised to overcome these problems. “Smart cities”, “smart nations,” and even a “smarter planet” were proposed. Various big IT companies decided to invest huge amounts of money into platforms designed to run the “cities of the future”. Fuelled by the upcoming Internet of Things, cities would be covered with plenty of sensors to automate them and thereby turn them into a technology-driven “paradise.” So far, however, these expectations have not been met.² Why?

Geoffrey West points out that cities cannot be run like companies.³ A company is oriented at maximizing profit, i.e. a single quantity, while a city must balance a lot of different goals and interests. This tends to make companies efficient, but vulnerable to mistakes. Cities are often less efficient, but more resilient. Driven by diverse interests, cities naturally do not put all eggs in one basket. This is why cities typically live longer than businesses, kingdoms, empires, and nation states.⁴

Importantly, cities are not just giant supply chains. They are also not huge entertainment parks, in which citizens consume premanufactured experiences. Instead, they are places of experimentation, learning, social interaction, creativity, innovation, and participation. Cities are places, in which diverse talents and perspectives come together, and collective intelligence emerges. Quality of life results, when many kinds of people can pursue their interests and unfold their talents while these activities inspire and catalyse each other. In other words, cities partly self-organize, based on a (co-)evolutionary dynamics.^5,6

While rapid urbanization comes with many problems, such as the overuse of resources, climate change and inequality,⁵ cities become ever more important, as they are motors of innovation.^3,5 Presently, more than half of humanity lives in cities, and the urban population is expected to increase to 68% by 2050. To meet the social, economic, and ecological challenges, innovation must be further accelerated, as the UN Agenda 2030 Sustainable Development Goals stress.

Given the digital revolution and the sustainability challenges, we now have to re-invent the way cities and human settlements are built and operated, and how cities can contribute to the solutions of humanity’s present and future existential problems. In the past, we had primarily two ways of addressing such issues: (1) nation-states (and their organization in the United Nations) and (2) global corporations. Both have not managed to deliver the necessary solutions on time, e.g. to problems such as climate change and lack of sustainability. Therefore, we propose a third way of addressing global problems: through networks of cities. ⁷ So, how to unleash the urban innovation engine?

CITY CHALLENGES

“City Olympics” or “City Challenges” could boost innovation on a cross-city level involving all stakeholders. They would be national, international or even global competitions to find innovative solutions to important challenges. Competitive disciplines could, for example, be the reduction of climate change, the development of new, energy-efficient systems, sustainability, resilience, social integration, and peace. The solutions would be publicly funded and should be Open Source (for example, under a Creative Commons license) in order to be reused and developed further by a multitude of actors in all cities i.e. by corporations, SMEs and spin-offs, researchers, NGOs and civil society. In this way, the potential of trends such as Open Source Movement, Hackathons, Fablabs, MakerSpaces, Gov Labs and Citizen Science would be raised to an entirely new level, creating the potential for civil society solutions. The new success principles would be collaborative practices such as co-learning, co-creation, combinatorial innovation, co-ordination, co-operation, co-evolution, and collective intelligence.

Increasing the role of cities and regions as drivers of innovation would allow innovative solutions and initiatives to be launched in a bottom-up way. All interested circles could contribute to City Challenges. Scientists and engineers would come up with new solutions and citizens would be invited to participate as well, e.g. through Citizen Science. Media would continuously feature the efforts and progress made in the various projects. Companies could try to sell better products and services. Politicians would mobilize the society. Overall, this would create a positive, playful and forward-looking spirit, which could largely promote the transformation towards a digital and sustainable society. In the short time available (remember that the UN wants to accomplish the sustainability goals by around 2030), the ecological transformation of our society can only succeed if the majority of our society is taken on board, and if everyone can participate and profit.

OPENSOURCING URBAN INNOVATIONS

Cities are the places where the engagement of citizens can have the greatest impact. The most liveable cities manage to create opportunities to unfold the talents of many different people and cultures and to catalyse fruitful interactions among them. Opportunities for participation and co-creation are key for success.

Alexandros Washburn⁸ said about the design process of New York City that he could not control anything, but influence everything; successful urban design required the right combination of top-down and bottom-up involvement. It is therefore essential that urban development involves all stakeholders including citizens. Vauban, a quarter of the city of Freiburg, Germany, is a good example for this. The city council encouraged the citizens to actively participate in land-use planning and city budgeting. Sustainability and new energy-saving technologies were a primary focus of the planning strategy. In two new districts (Rieselfeld and Vauban), self-built and community architecture was created, which led to urban environments conceived and designed by future inhabitants according to their own vision. Now, Freiburg counts as benchmark city. Its concepts of sustainable urban planning and community participation are widely used by other cities all over the world.

So far, most urban planning professionals do not pay much attention to long-term involvement of citizens in urban development. With the ubiquity of information and communication technologies, our cities are getting smarter, but not automatically more inclusive, just, and democratic. The Citizen Score, a surveillance-based approach to control the behaviours of people, shows how easily technological progress may lead to technological totalitarianism. In the private sector as well, global corporations, geared towards profit, can turn into threats of democracy and human rights. When services are free, people are the product, data can fall into the wrong hands, and human dignity, autonomy, and freedom will be compromised. In data-rich societies, where people are measured and watched, profiled and targeted, this problem is quite significant. If cities of the future were run like businesses, based on surveillance, driven by data and controlled by algorithms, liberty, democracy, and human rights might quickly erode.

The application of open source principles to the co-creation of urban environments could overcome these problems by supporting active participation, technological pluralism and diversity. Thereby, it would also avoid technological lock-ins and dead-ends. The open source movement, which started with opening software (see the example of GitHub) now promotes the co-production of open content (Wikipedia, OpenStreetMap), open hardware (3D-printer RepRap), and even open architecture (WikiHouse). Open Source Urbanism would be the next logical step of this open source trend.

In 2011, Saskia Sassen wrote: “I see in Open Source a DNA that resonates strongly with how people make the city theirs or urbanize what might be an individual initiative. And yet, it stays so far away from the city. I think that it will require making. We need to push this urbanizing of technologies to strengthen horizontal practices and initiatives.”⁴

Yochai Benkler argues that open source projects indicate the beginning of a social, technological, organizational and economic transformation of the society towards a new mode of production.⁹ This new mode, called commons-based peer production, is a collective activity of volunteers, usually coordinated via the Internet, producing free-to-use knowledge. Open Source Urbanism, as a new way of urban development, would therefore build on concepts such as Open Source Innovation and Commons-Based Peer Production.

In fact, citizens are keen to be not just consumers, but co-producers of their urban habitats. Some of them already experiment with open-sourcing urban design by collecting, improving, and sharing their Do-It-Yourself design blueprints and manuals on the Internet. The “Nation of Makers” initiative promotes community-driven design, prototyping, and fabrication as well in order to solve local and global challenges by improving lives in local communities around the planet.

Such examples are presently still rare and dispersed, and, therefore, not yet able to shift cities towards more inclusive urban development on a global scale. For this, one would need a socio-technical platform to consolidate and strengthen the nascent movement. Such a platform could promote the exchange of best practices and solutions to frequently occurring problems. The results would be digital commons designed to satisfy citizens’ needs¹⁰.

The proposed approach pushes for a new paradigm of globalisation, which one may call “glocalisation”. It would be based on thinking global, but acting local (and diverse), on experimentation, learning from each other, and mutual support. The approach would be scalable. It would be more diverse and less vulnerable to disruptions. It would promote innovation and collective intelligence, while being compatible with privacy, freedom, participation, democracy, and a high quality of life. If cities would open up and engage in co-creation and sharing, they would quickly become more innovative and efficient. Open Source Urbanism could take our cities and societies to an entirely new level and also help to create better living conditions in developing countries and regions suffering from war more quickly.  

References

1.        Sennett, R. Building and Dwelling: Ethics for the City. (Farrar, Straus and Giroux, 2018).

2.        Hugel, S. & Hoare, T. Disrupting cities through technology, Wilton Park. (2016).

3.        West, G. Scale: The Universal Laws of Growth, Innovation, Sustainability, and the Pace of Life in Organisms, Cities, Economies, and Companies. (Penguin, 2017).

4.        Sassen, S. Open Source Urbanism. Domus (2011). Available at: http://www.domusweb.it/en/op-ed/2011/06/29/open-source-urbanism.html. (Accessed: 16th November 2016)

5.        Bettencourt, L. M. A. & West, G. A unified theory of urban living. Nature 467, 912–913 (2010).

6.        Batty, M. Cities and complexity: understanding cities with cellular automata, agent-based models, and fractals. (The MIT press, 2007).

7.        Barber, B. R. If mayors ruled the world: Dysfunctional nations, rising cities. (Yale University Press, 2013).

8.        Washburn, A. The nature of urban design: A New York perspective on resilience. (Island Press, 2013).

9.        Benkler, Y. Freedom in the Commons: Towards a Political Economy of Information. Duke Law J. 52, 1245–1276 (2003).

10.      Schrijver, L. in Handbook of Ethics, Values, and Technological Design: Sources, Theory, Values and Application Domains (eds. van den Hoven, J., Vermaas, P. E. & van de Poel, I.) 589–611 (Springer Netherlands, 2015). doi:10.1007/978-94-007-6970-0_22

On the Use of Big Data and AI for Health

Pitfalls of Big Data Analytics

High-precision medicine requires reliable decisions whom to treat best in what way, when and with what dose of what medicine, ideally even before a disease breaks out. This challenge, however, can only be met with large amounts of personal and/or group-specific data, which may be extremely sensitive, as such data may be used against the interest of the patients (e.g. in the interest of profit maximization). Consequently, there are plenty of technical, scientific, ethical and political challenges.

This situation makes it particularly important to protect personal data from misuse by means of cybersecurity, to ensure a professional use of the data, and to implement suitable measures to achieve a maximum level of human dignity (including informational self-determination).

In the past, empirical and experimental analyses have often been suffering from lack of data or small amounts of data. In many areas, including medical studies, this has changed, or is about to change. Big Data is, therefore, promising to overcome some common limitations of previous medical treatments, which were often not personalized, imprecise, ineffective and connected with many side effects.

In the early days of Big Data, people expected to have found a general purpose tool, something like a holy grail. It was believed that, if one had just enough data, data quantity would turn into data quality; the truth would basically reveal itself. This idea is probably best expressed by a quote by Chris Anderson, who – back in 2008 – predicted “the end of theory” and wrote in the Wired Magazine: “The data deluge makes the scientific method obsolete.”

Along these lines it was claimed that it would now be possible to predict, or at least to “nowcast” the flu from Google searches, as reflected by the platform Google Flu Trends. The company 23andMe offered to identify ethnic origin, phenotype, and likely diseases. Angelina Jolie said “knowledge is power” and had her breasts removed, because her genetic test identified a high chance she would get breast cancer.

Later on, Google Flu Trends was closed down, doctors warned that Angelina Jolie should not be taken as example, and 23andMe’s genetic test was temporally taken off the market by the health authority. How could this happen? Google searches were not anymore a reliable measurement instrument, as Google had started to manipulate people with suggestions (both through the autocomplete function and by means of personalized advertisement). Regarding attempts to predict diseases by means of genetic data, it was discovered that some people were doing very well, even though they were predicted to be very ill. Moreover, predictions were sometimes quite sensitive to adding or subtracting data points, to the choice of the Big Data algorithm, or (in some cases) even to the hardware used for the analysis.

Generally, it was thought that – the more data one would have the more accurate the implications of data analyses would be. However, the analyses often took correlations for causation, and they were not checking for statistical significance – in many cases, it was not even clear what the appropriate null hypothesis was. So, in many cases, Big Data analytics was initially not compatible with established statistical and medical standards.

In fact, the more data one has, the higher the probability to find patterns in the data just by chance. These patterns will often be not meaningful or significant. Spurious correlations are a well-known example for this problem. These are correlations that do not reflect a causal relationship, or where a third factor causes two effects to correlate, where neither effect influences the other. In such cases, increasing or decreasing the measured variables would not have the expected effect. It could even be counterproductive. Careful causality analysis (by concepts such as Granger causality) are, therefore, absolutely required.

Another problem concerns undesirable discrimination. Suppose a health insurance wants to incentivize certain kinds of “healthy” diets – by reducing tariffs for people who eat more salad and less meat, for example. As a side effect, it would then be likely that men will pay different tariffs from women, and Christians, Jews, and Muslims would on average pay different tariffs as well, just because of their different religious and cultural traditions. Such effects are considered discriminatory and need to be avoided. If one, furthermore, wants to avoid discrimination based on age, sexual orientation and other features that should not be discriminated against, Big Data analytics becomes a quite sophisticated challenge.

Last but not least, even Big Data analytics will produce errors of first kind and of second kind, i.e. false alarms and alarms that don’t go off. This is a problem for many medical tests. Say, a medical test costs x and a correct diagnosis creates a benefit of y, while a wrong one will cause a damage of z. Moreover, assume that that the test is correct with probability p and incorrect with probability (1-p). Then, the overall utility of the test is u = – x + p*y – (1-p)*z, which might be neutral or even negative, depending on the impact of wrong diagnoses. For example, false negatives are an issue for many kinds of cancer, and it is therefore sometimes advised, not to test the entire population.

In conclusion, the scientific method is absolutely indispensable to make sense of Big Data, i.e. to refine raw data into reliable information and useful knowledge. Hence, Big Data is not the end of theory, but rather the beginning.

A good example to illustrate this is the example of flu prediction. When the spatio-temporal spreading of the flu is studied, one will often find a wide scattering of the data and a low predictive power. This is related to the fact that the spreading of the flu is related to air travel. However, it is possible to use data of the passenger volumes of air travel to define an effective distance between cities, where cities with high mutual passenger flows are located next to each other. In this effective distance representation, the spreading pattern becomes circular and predictable. This approach makes it possible to identify the likely city in which a new disease emerged and to forecast the likely order in which cities will be suffering from the flu. Hence, it is possible to take proactive measures to fight the disease more effectively.

Pitfalls of Machine Learning and Artificial Intelligence

With the rise of machine learning methods, new hopes emerged that the previously mentioned problems could be overcome with Artificial Intelligence (AI). The expectation was that, AI systems would sooner or later become superintelligent and capable of performing any task better than humans, at least any specialized task.

In fact, AI systems are now capable of performing many diagnoses more reliably than doctors, e.g. diagnoses of certain kinds of cancer. Such applications can certainly be of tremendous use.

However, AI systems will make errors, too, just perhaps with lower frequency. So, decisions or suggestions of AI systems must be critically questioned, particularly when a decision may have large-scale impact, i.e. when a single mistake can potentially create large damage. This is necessary also because of a serious weakness of most of today’s AI systems: they do not explain how they come to their conclusions. For example, they do not tell us what is the likelihood that the suggestion is based on a spurious correlation. In fact, if AI systems turn correlations into laws (as cybernetic control systems or autonomous systems may do), this could eliminate important freedoms of decision-making.

Last but not least, it has been found that not only humans, but also AI systems can be manipulated. Moreover, intelligent machines are not necessarily objective and fair: they may discriminate people. For example, it has been shown that people of color and women are potential victims of such discrimination, in part because AI systems are typically trained with biased, historical data. So, machine bias is a frequent, undesired side effect and it is a serious risk of machine learning, which must be tested for and properly counter-acted.

Künstliche Intelligenz kann eine Chance für uns alle sein

Von Dirk Helbing 
(ETH Zürich, TU Delft, Complexity Science Hub Vienna)
 
Es war lange ein Traum des Silicon Valleys, Künstliche Intelligenz (KI) zu bauen, die intelligenter als Menschen ist und die Probleme löst, die uns Menschen über den Kopf gewachsen sind. KI hätte unsere menschlichen Fehler nicht, dachte man. Sie wäre objektiv, fair, und unemotional, könnte viel mehr Wissen überschauen, schneller entscheiden und aus Daten lernen, die in der ganzen Welt gesammelt werden. Städte könnte man mit Mess-Sensoren versehen und automatisieren. Am Ende stünde eine Smarte Gesellschaft, die sich datengetrieben und algorithmen-gesteuert optimal entwickelt. Wir müssten nur tun, was uns das Smartphone sagt. Verhaltenssteuerung durch personalisierte Information und den berühmtberüchtigten chinesischen Citizenscore, ein Punktekonto für das Wohlverhalten des Bürgers, würde für die optimale Gesellschaftssteuerung sorgen. Inzwischen ist da vielerorts Ernüchterung eingekehrt. Was einst als Utopie begann, wird heute oft als Alptraum angesehen.

Damit treten wir in eine neue Phase der Digitalisierung ein. Die Karten werden neu gemischt. Europa hat die Chance, eigene Impulse zu setzen und damit Weltmarktführer zu werden – durch Künstliche Intelligenzsysteme, die Menschen nicht überwachen und kontrollieren, sondern die Menschen befähigen und kreative Aktivitäten koordinieren. Die Rede ist nun vom „werte-sensitiven Design“. Gemeint ist: wir sollten unsere verfassungsrechtlichen, sozialen, ökologischen und kulturellen Werte in die intelligenten Informationsplattformen einbauen, damit sie uns dabei unterstützen, unsere gesellschaftlichen Ziele zu erreichen, aber Freiräume für Kreativität und Innovation lassen.

Wenn es um demokratische Werte geht, so sind etwa die folgenden Aspekte von Bedeutung: Menschenrechte und Menschenwürde, Freiheit, (informationelle) Selbstbestimmung, Pluralismus, Minderheitenschutz, Gewaltenteilung, Checks and Balances, Mitwirkungsmöglichkeiten, Transparenz, Fairness, Gerechtigkeit, Legitimität, anonyme und gleiche Stimmrechte und Privatsphäre im Sinne von Schutz vor Exponierung und Missbrauch einerseits, andererseits im Sinne eines Rechts, in Ruhe gelassen zu werden.

Im globalen Miteinander scheinen überdies folgende Werte eine vielversprechende Basis für eine erfolgreiche und friedliche, vernetze Informationsgesellschaft zu sein: Vielfalt, Respekt, Partizipationschancen, Selbstbestimmung, Verantwortung, Qualität, Awareness, Fairness, Schutz, Resilienz, Nachhaltigkeit und Compliance.

Es ist nicht leicht, diese Eigenschaften in Informationssysteme einzubauen, aber wir können es lernen. Wir können KI-Systeme bauen, welche die Welt und uns alle voranbringen, vorausgesetzt es gibt einen breiten und fairen Zugang zu den Potenzialen dieser Systeme. Stellen Sie sich vor, die KI würde Ihnen nicht sagen, was Sie tun sollen, sondern sie würde Ihnen dabei helfen, Ihre eigenen Talente zu entfalten und Ihre Ziele zu erreichen, und zwar umso mehr, je mehr sie (auch) anderen helfen – sozusagen ein Geist aus der Flasche, der Gutes tut, der uns hilft, uns selbst und anderen zu helfen.

Was sich heute noch wie Utopie oder Science Fiction anhört – schon bald könnte es Realität sein. KI ist eine Chance für die Wirtschaft, für Europa und uns alle, wenn wir nur lernen damit umzugehen – damit es nicht ausgeht wie mit Goethe’s Zauberlehrling. Die Enquete-Kommission „Künstliche Intelligenz – gesellschaftliche Verantwortung und wirtschaftliche Potenziale“ hat jetzt die Chance, die Weichen für eine vielversprechende, bessere Zukunft zu stellen.