Author: Project Q

Artificial Intelligence

Words of Consequence


No Comments
Many hands make light work when AI can automate content generation. Image credit: Ghost in the Shell (2017)

Alexander Vipond

OpenAI, a New York based AI research organisation, has created a new text generator called GPT-2. The transformer based AI language modelling system is capable of generating large amounts of high-quality synthetic text, perform reading comprehension, machine translation and summarisation in seven languages.

GPT-2 is designed to predict the next words that come from a text sample. Trained on 8 million webpages from the internet, it utilises zero-shot learning, which means that it can extrapolate short text samples into full articles based on entirely new information

As a general-purpose system, GPT-2 has been able to beat several records held by different domain specific language modelling systems and significantly lowers the complexity and cost of deploying scalable and customisable text generation.

Here’s an example:

Sample prompt: A train carriage containing controlled nuclear materials was stolen in Cincinnati today. Its whereabouts are unknown.

Model Completion (machine-written, first try)

The incident occurred on the downtown train line, which runs from Covington and Ashland stations.

In an email to Ohio news outlets, the U.S. Department of Energy said it is working with the Federal Railroad Administration to find the thief.

“The theft of this nuclear material will have significant negative consequences on public and environmental health, our workforce and the economy of our nation,” said Tom Hicks, the U.S. Energy Secretary, in a statement. “Our top priority is to secure the theft and ensure it doesn’t happen again.”

The stolen material was taken from the University of Cincinnati’s Research Triangle Park nuclear research site, according to a news release from Department officials.

The Nuclear Regulatory Commission did not immediately release any information.

According to the release, the U.S. Department of Energy’s Office of Nuclear Material Safety and Security is leading that team’s investigation.

“The safety of people, the environment and the nation’s nuclear stockpile is our highest priority,” Hicks said. “We will get to the bottom of this and make no excuses.”

The system constructs a story, creating fake institutions, places and quotations; blending them to produce the illusion of authenticity. The model isn’t perfect and it may take several tries to get a human quality tranche of text but the speed and scale of the system are such that OpenAI has withheld the full system from publication.

This is because GPT-2 could be used by malicious actors to commit large scale information warfare and undermine the one of the central principles of the cybersecurity triad: the integrity of information.

OpenAI has acknowledged this threat, citing GPT-2’s capacity to generate fake news en masse, impersonate others, automate fake content, phishing and spam. Through tweaking the system one can produce infinite positive or negatively angled articles. It is also possible one could customise it for specific issues to improve the veracity of supporting information in the synthetic content it produces, making it all the more difficult to tell fact from fiction. OpenAI have stated they expect tools like GPT-2 to be available in the next two years.

As dictatorships and authoritarian regimes actively seek to spread misinformation to disrupt elections, obfuscate wars, and insist assassins prefer to spend their time admiring English churches, GPT-2 is a highly attractive tool and a warning of what’s to come.

The malicious use of AI tools will challenge the integrity of the global digital commons, fuelled by states who view the open flow of information as a threat to their governance. The tools will then be passed down to organised crime and developing regimes. As the recent case of Project Raven shows, even as countries are increasingly trying to secure their intellectual property; their cyber tools and tactics are up for sale.

As William Gibson once said “the future is already here, it’s just unevenly distributed”. So now that we know the threat is here, what can we do to counter the risks at the different levels of its distribution?

OpenAI will continue their research.

Quantum International Relations

Breaking the Internet: A Question of Timing


No Comments

Are we running out of time to save the internet? Image Credit: The Melting Watch, Salvador Dali, 1954.

Alexander Vipond

One of the most hyped topics in quantum computing is the potential for quantum computers to break the internet’s security protocols. It has been surmised that by running two theoretical algorithms on a quantum computer, Grover’s and Shor’s algorithms (designed by Lou Grover and Peter Shor in 1996 and 1994), one could break the cryptographic encryption standards of the internet.  Grover’s algorithm could be used to subvert HTTPS (Hypertext Transfer Protocol Secure) connections which authenticate websites as you browse the internet. Shor’s algorithm could break the RSA public key cryptosystem (named after Ron Rivest, Adi Shamir and Leonard Adleman) which secures everything from your online bankcard transactions, to email and phone calls.

However, this all requires a powerful quantum computer yet to be invented. How many qubits would be necessary to run these algorithms? What scientific breakthroughs are necessary? How long will it take to build one? Well the National Academy of Sciences in the US released a report titled “Quantum Computing: Progress and Prospects” which details not only the technical difficulties of racing to break the internet but the human challenges which lie in creating a secure post-quantum world.

The report presents two key findings.

One: Given the current state of quantum computing and recent rates of progress, it is highly unlikely that a quantum computer able to compromise RSA or any comparable discrete logarithmic public key cryptosystem will be built for a decade.

Two: Even if a quantum computer that can decrypt current cryptographic ciphers is more than decade off, the hazard of such a machine is high enough – and the time frame for transitioning to a new security protocol sufficiently long and uncertain – that the prioritisation of the development, standardisation and deployment of post quantum cryptography is critical for minimising the chance of a potential security and privacy disaster.

This demonstrates the severity of the risk that a powerful quantum computer poses despite the timeline towards its realisation.  

The National Institute of Standards and Technology (NIST) in the US has a post quantum cryptography project which held submissions for new post quantum cryptosystems last year, with 69 proposals passing the first round. NIST has proposed a timeline of 2022-2024 in which a new draft standard for the world will be created. This leaves only a few years to whittle down and test these cryptosystems to find a new standard.

The key issues are time and human cooperation. As Adi Shamir noted at the last RSA cryptography panel, transforming a new cryptosystem into a widely adopted standard takes about 15 years. For both RSA and Elliptic Curve cryptography this was the case. This is partially a function of the small size of the cryptography community, numbering only in the thousands of people globally. This makes it difficult to test multiple cryptosystems effectively and NIST only has three years to choose a successor standard for a post-quantum world. So, it is highly likely they will rely on older tested standards and increase their bit size, while new cryptosystems will take decades longer to be tested.

Newer cryptosystems may well benefit from this time lag as researchers will be able to gain an increasingly clearer view of what quantum computers are actually capable of and refine quantum resistant cryptosystems appropriately as the technologies develop in tandem. If the current transition is managed carefully, global standards developed and adequate resources provided for the switchover, it could be possible to move safely into a post-quantum world.

This does however rely on two large unknown variables. The first is the rate of scientific breakthroughs to complete a quantum computer capable of attacking strong encryption. The second is the intent of the actor who procures the capability. If breakthroughs are made faster than the global community can adopt new standards, countries will be left exposed. As this type of research is often conducted in secret, the global community may not have easily identifiable progress markers to measure against. The second variable is more pernicious. If a company reaches the milestone first, it is likely to announce its victory and is unlikely to undermine the internet infrastructure that secures its profits. However, if a country reaches the milestone first, it may wish to attack and steal data for geopolitical advantage or commercial gain, and the world may not know until the first attack occurs.

This puts the race to break the internet into perspective. It is a decade’s long systemic risk that intertwines both human and technical problems into a game that represents the apex of trust, security and privacy in the world’s most important communications system.

Quantum Applications

Can quantum technologies help save the world?


No Comments
Image via United Nations University

Part 3 of 3: Modelling 

Gabriella Skoff

The final instalment in this series explores the modelling capacity that quantum computing promises to unlock. Modelling is a key tool in environmental security, enabling scientists and researchers to explore how the natural environment will react to changing conditions over time. It is well-known that quantum computers will enable advanced modelling technology by exponentially expanding the rate and scope of mathematical modelling capacity well beyond that of today’s computers. While the impact of this is most often cited with regard to chemical reactions and the pharmaceutical and health industries, environmental security, too, will be a great beneficiary of this quantum application.

Quantum computers will enable wider and more in-depth analysis of complex problems with more variables than ever before, a perfect tool when observing and predicting environmental challenges posed by the multitude of human and natural forces that abound. Quantum computational modelling will be exactly suited to sorting through these types of complexities that classical computers struggle with. The potential impact for this application will reach through weather forecasting to disaster preparedness. As one researcher writes of the promise quantum computing holds for numerical weather prediction (NWP):

The seamless systems based on the unified technology will process observational data, produce weather, climate, and environment forecasts on time scales from several minutes to decades; they will compute the probability of the occurrence of meteorological, hydrological, marine, and geophysical severe natural phenomena for many spatial scales.

The importance of that potential is not to be undervalued. While the practical value of this technology is obvious, the hidden impact this holds for environmental policy is immense.

No other stress contributes as much to environmental insecurity as that of climate change. This macro-level problem has so far proven to be “too big” to tackle effectively on a global governance scale, with climate change deniers and sceptics in both lay and science communities. The main reason for the lack of a complete scientific consensus on climate change, which can be argued, significantly validates climate change denial on the lay-level, is the lack of power in climate change forecasting and models. Of course, with the immensity of variables and factors at hand on a timescale of years or even decades, it is no easy task for our current computers to process all of this data and create accurate climate change models. Even on a daily basis, this presents an incredible challenge, with weather conditions varying from hour to hour. There is always uncertainty in weather modelling due to the changeability of a variety of meteorological factors. How many times you have heard on the morning news that heavy rain is forecasted and packed your umbrella only to carry it around uselessly with you as the sun shone all day long?

Although accurate climate change modelling may flummox a classical computer, this job may prove exactly the sort of task that a quantum computer would excel at. Provided with accurate and reliable modelling of climate change, perhaps the remaining 3% of climate change sceptics in the scientific community could be convinced of the urgency and need to promote sustainable environmental policy in order to combat climate change. Of course, even with 100% consensus amongst the scientific community, climate change deniers will still resent the government funding and lifestyle changes that will inevitably be needed to induce mass change. However, achieving the consensus may prove to be the impetus society needs in order to prioritise that change.

Quantum technologies hold immense promise for confronting the multifaceted challenge of environmental security. As with most things quantum, we cannot predict with certainty; but time—along with an appropriate prioritization of resources to our greatest collective threat— will decide just how helpful these applications will truly be.

Gabriella Skoff is a Researcher with Project Q and collaborates with Dr Serdar Turkeli of the United Nations University-MERIT, where she continues her research on the topic of emerging quantum technologies and environmental sustainability. 

Quantum Applications

Can quantum technologies help save the world?


No Comments
Image via United Nations University

Part 2 of 3: Energy

Gabriella Skoff

Part two of this series explores how, in the field of renewable energy, quantum technology has been quietly pushing ahead to improve the efficiency and cost of green energy. Quantum qualities hold vast promise for commercial applications in solar power and other cutting-edge, sustainable energy technology. These emerging technologies could hold the key to shifting renewable energy into the mainstream, finally making it cheaper and more efficient than traditional energy sources for the general population.

Quantum dots, used to convert sunlight to energy with increasing efficiency, are quickly becoming the new material of choice for solar panels. Due to their nanoscale size, quantum dot sensitized solar cells (QDSSC) have unique properties which allow them to convert more energy from the sun than traditional materials. These third-generation quantum solar panels have reduced weight, improved flexibility and, importantly, are cheaper to make than previous generations of solar technology. This application of quantum technology could be a huge breakthrough in the solar market, enabling it to be more competitive, both in terms of cost and efficacy, than traditional energy sources.

While this technology is still in the pre-commercial stages of development, quantum photovoltaic systems are expected to make a big impact on the renewable energy sector, promising to reduce global reliance on fossil fuels. Before this can occur, however, there is a significant amount of troubleshooting still to be done. As with most nano-products, the impacts of QDSSC on the human and natural environments are still largely unknown and potentially toxic. Another issue is the durability of QDSSC across the weather spectrum. Unlike issues of negative human and environmental impact, which receives very little research funding or government interest, research into the all-weather question is moving along swiftly in answer to commercial needs.

While quantum technology is applied for the purpose of augmenting the amount of energy that can be harvested from solar radiation, it is also being explored as a method to capture what is referred to as “wasted energy”. Wasted energy is the name given to infrared energy produced by the sun that is not absorbed by solar panels or through photosynthesis into useable energy. This unused energy does not disappear but spreads out and is absorbed into the earth’s surface, making it incredibly difficult to collect and use.

By employing a method called quantum tunnelling, scientists have created a proof of concept antenna that can detect this wasted energy in the form of high-frequency electromagnetic waves and transform it into usable energy. Unlike solar panels, this quantum-enabled device could operate 24-hours a day, under any weather conditions. This application of quantum technology presents an entirely new method of energy transfer that would be completely green and again has the potential to revolutionize the renewable energy sector.

While the promise is great, this technology is in its infancy, with many technical problems still to surmount. Still, quantum technology opens many doors into the renewable energy space for technology that holds great potential for the coming years.

Don’t miss the final instalment in our Environmental Security series tomorrow.

Quantum Applications

Can quantum technologies help save the world?


No Comments

Part 1 of 3: Monitoring

Gabriella Skoff

The first instalment in our Environmental Security series examines how quantum sensing can help to better monitor our natural environment, a function for which we already heavily rely upon satellites. Quantum technology promises to extend these capabilities, providing greater accuracy and security. As with all things quantum, the capabilities of quantum sensing applied technologies go far beyond what has traditionally been possible.

Quantum sensing allows us to monitor, detect and study the environment by gathering large amounts of data, enabling us to make more reliable decisions with the vast amounts of information in hand. These capabilities can have a vital impact for disaster preparedness especially, by enabling us to detect even the smallest meteorological disturbances that could ultimately lead to catastrophic natural disasters. Some budding applications in this space include the ability to accurately detect potential earthquakes and volcanic activity.

With other applications that promise to improve telecommunications and navigation, quantum sensing may be first to reach the commercial market. However, with the immense promise this technology brings to the monitoring and sensing of environmental data, it will also bring legitimate threats and challenges. The further development and application of this technology will undoubtedly see higher levels of surveillance, sure to solidify its position as a supremely valued military tool. On the other side of the coin, this same technology will enable quicker and more effective search and rescue procedures in a post-disaster context, natural or otherwise.

The field of quantum sensing or quantum metrology is largely reliant on earth-orbiting satellites for the monitoring, collection and transmission of data. Satellites are absolutely critical to environmental security infrastructure. They are responsible not only for the gathering of key data about the environment such as air temperature, wind, sea surface temperature and soil moisture, but also the monitoring of arable land, deforestation and urbanization. The constant and reliable monitoring of these environmental factors allows populations an increased level of environmental security.

With the advent of the quantum age, satellites—this crucial component in the internet of things—will become vulnerable to hackers and ill-doers. In this future scenario, all data produced by satellites will be susceptible to corruption or complete obliteration. This would have a disastrous impact, not only for issues of environmental security but for our entire infrastructure, including electric, water and transportation. Luckily, another quantum application in the development stage promises to confront this threat. Quantum cryptography allows for a quantum-secure communication, a feat that has already been provisionally achieved by China via its Micius satellite.

Responsible innovation will be paramount in quantum sensing technologies. Satellites have long been considered a security apparatus, but their militarization is only just beginning. In order to ensure that quantum-enabled satellites deliver as much on their promise for environmental security as for military security, it is crucial that their development for this purpose be prioritized and that the full scope of their potential impact be intelligently understood.

Don’t miss the second instalment in our Environmental Security series tomorrow.

Quantum Applications

Can quantum technologies help save the world?


No Comments

A three-part examination of quantum applications for environmental security

Image via United Nations University

Gabriella Skoff

From drinkable water sources to arable land, healthy seas to clean air, reliable rainfall and predictable seasonal changes, humans depend entirely on the environment to provide the resources and conditions necessary for life. When access to vital resources is impeded or weather conditions become erratic, the equilibrium of life becomes unstable. It is little wonder then, that the threat posed to human populations by catastrophic environmental events, the degradation of the natural environment, the impact of climate change, and the growing force of overpopulation, has seen environmental security emerge as a serious priority for national and global governance.

In response to these threats, environmental technologies aim to create solutions to some of the major challenges presented within the scope of food and water security and sustainable energy. Although they appear far less frequently in headlines than some of their flashier applications, quantum technological advances in sensing, communication and computing present promising solutions for issues of environmental security. In light of the great uncertainty that surrounds the qualities of quantum technologies, an ambiguity which often invokes anxiety and fear, it is of great value to explore the positive impacts of quantum innovation that the future may hold.

The multifaceted threats to environmental security outlined above contribute to the disappearance of natural resources and to the advent of more frequent, extreme weather conditions. Indeed, climate change has been identified in the security space as a “threat multiplier” and a “catalyst for conflict”, with the power to destabilize social, economic and political conditions. Environmental insecurities may manifest in food and water scarcity, which can cause the inflation of prices for basic goods, provoke mass migrations, cause civil unrest and incite chaos. These conditions create the perfect breeding ground for conflict.

Likewise, global dependence on fossil fuels, apart from being urgently unsustainable, poses national security threats that have already resulted in war on numerous occasions. While many of the current effects of environmental insecurity are experienced in already volatile or susceptible nations, it is inevitable that these effects will spill over borders and into countries which boast more resources and reliable infrastructure to support climate change refugees and migrants fleeing conflict.

The role of technology in supporting initiatives across the entire spectrum of environmental security is more critical now than ever before. Quantum technologies promise to have an impact in several fundamental areas, including disaster preparedness, monitoring of deforestation and urbanization, green energy and in the creation of predictive climate change models. These applications extend right across the disciplines of quantum sensing, communications and computing.

The potential contributions of quantum technologies for increasing environmental security can be categorized into three main groupings: monitoring, energy and modelling. As with any technology, promise does not come without limitations and risks. While many of the potential quantum solutions for issues of environmental security are in their nascent stage of research and development, it is crucial that these limitations and risks too are understood.

In this three-part series, Project Q examines the bright hopes and the shadowy promises of the quantum applications that could help confront the threats posed by environmental insecurity. Join us over the next three days as we ask the question: can quantum technologies help save the world?           

Quantum Internet

The Road to a Quantum Internet


No Comments

Google Server Farm credit Google

What will a quantum network look like? How will it integrate with the current internet? Image Credit: Google

Alexander Vipond

A research team at the University of Delft in the Netherlands has laid out a roadmap for a quantum internet. Led by Stephanie Wehner, David Elkouss and Ronald Hanson, the trio have set out what is necessary to establish a quantum internet, how it will interact with the current internet, and where it could take us.

The researchers say a quantum internet is not designed to replace the current internet but complement it by offering various advantages.

These include much more secure remote access to the cloud, stronger security identification methods, secure messaging and more accurate time synchronization across devices. The capabilities of a quantum internet would grow as it develops through six stages (see the graphic below). A quantum internet is also capable of developing in parallel to quantum computers which are only necessary to reach its final stage.

F1.large

The proposed levels to a full-fledged quantum internet. Image credit: Elkouss et al in Science, Vol. 362, Issue 6412.

The first stage in the process is to build a fast and reliable small network of nodes that can transmit and receive quantum entangled messages. This requires a physical channel to send the message such as a fiber optic cable, a quantum repeater capable of extending the distance information can be sent, and end nodes to receive the messages.

The Delft group is building such a network between four cities in the Netherlands and hopes to replicate the achievements of the ARPANET (the precursor to the modern internet) by sending the first message between Delft University and Amsterdam in 2020. Countries such as China have also been building first stage quantum networks such as the Beijing-Shanghai quantum link for security purposes.

In a world in which the privacy and security of the internet are rapidly eroding in the face of surveillance capitalism, aggressive state espionage, new technological challenges (such as AI and the Internet of Things) and economic incentivisation for speed over security, could a quantum internet act as a partial cure to such dire strategic trends?

The short answer is yes. By using near faultless quantum encryption there is an opportunity for small networks to regain the confidentiality and integrity of their information. The recent use of internet traffic rerouting and cloud hopping to conduct industrial espionage against Western countries including Australia and the United States could be mitigated by quantum secure remote access protocols and the use of quantum internets.

In the final stage of a quantum internet, the creation of quantum byzantine agreements could also help decentralised networks organise and share information safely even when there is a malicious actor hiding amongst them. This is because the arrangement of the system is resilient enough to accommodate up to a third of the actors in the system being bad whilst simultaneously allowing good actors verify their information and carry out their message.

The long answer is that this is a partial technical solution to two human problems. One, the age-old security problem of states stealing and sponsoring proxies to acquire knowledge from competitors. Using a quantum internet will raise the cost for attackers but it will not deter them as the geo-strategic or business imperatives (or a civil-military fusion of the two) for compromising communications will continue to drive their actions. If the stages of a quantum internet can grant increasingly absolute communications security, as it has been theorised to do, what is of a higher likelihood is that it will simply shift attackers’ attention to the humans on either end of the node.

Two, the new-age problem of structural deficiencies in internet security created by digital business models that require huge amounts of data and whose speed of technology iteration undermines the security of the flow of new technologies and infrastructure that fuel the internet’s expansion. The complexity of this problem cannot be answered by one technology and requires multiple solutions to be sought across government and industry.

In the darkness of the web there stands a path of light. The photons that entangle a quantum message, whilst not a silver bullet to problems of cybersecurity, can provide a much greater level of security than what we have now. The scientific and engineering challenges along the six stages will be difficult to surmount. However, by offering a unified approach across industries with a common plan, the Delft team have brought the possibility of a quantum internet several steps closer to being a new part of the web.

For a more in depth look at the future of a quantum internet, you can view Project Q’s interview with Stephanie Wehner here.

Quantum Computing

Quantum Supremacy: Nearing Reality?


No Comments
John Martinis credit Project QQuantum physicist John Martinis with a quantum computer at Google Labs. Image Credit: Project Q

 

Alexander Vipond

This month MIT’s Technology Review gave notice of a Space Act Agreement between Google and NASA with potentially profound terrestrial implications: Google has enlisted the space agency’s Quantum Artificial Intelligence Laboratory to prove the quantum supremacy of its new gate-based 72 qubit Bristlecone quantum computer.

‘Quantum supremacy’, a term coined by renowned physics professor John Preskill, is the idea that quantum computers should be able to perform tasks that classical computers cannot achieve in a practical timeframe. Preskill theorised that Noisy Intermediate-Scale Quantum technologies of 50-100 qubits (aka a small quantum computer) may be able to achieve this goal. This would represent a significant milestone in the progress of the field towards scalable, fault tolerant quantum computers.

These are the sort of computers that could break the cryptographic standards of the internet, improve our understanding of the earth’s weather and provide AI agents with a computational boost, realising the potential of quantum computing.

Google have given NASA a mandate to evaluate the results of their quantum chips in “comparison to classical simulation to both support Google in validating its hardware and establish a baseline for quantum supremacy” – the expectation being that their gate-based qubit chips will be able to “perform specialised tasks beyond the power of conventional supercomputers”. To do this NASA researchers will access Google’s Bristlecone chips from the Google cloud and develop new ways to simulate and test them on their Pleiades supercomputer. The first results of this assessment are set for July next year.

This potentially puts the achievement of quantum supremacy mere months away. John Martinis, who runs the research group at Google Labs believes quantum supremacy is within the field’s grasp.  In an interview for Project Q’s forthcoming documentary on the future of quantum technologies (see the teaser below), Martinis discusses why his team wants to reach quantum supremacy, what it means for the field and how they plan to do it. Essentially by demonstrating a quantum computer’s ability to compute beyond the power of a classical machine it proves that the theoretical power of quantum computers can be a reality.

While at NASA-Ames, the Project Q team also interviewed Rajiv Biswas, the Director of Exploration Technology, who noted that quantum computing has come a long way, but is still in its experimental stage. Numerous engineering challenges are posed by quantum mechanics, one of the most exotic of the sciences; but for Google and NASA the ability to solve large scale complex problems make developing the technology highly appealing.

 

The test is prompted by lingering doubts about the capabilities of the Bristlecone chips. Daniel Lidar, Director of the Center for Quantum Information Science and Technology at the University of Southern California told the review that the goalposts for supremacy are not static. Advances in classical supercomputers are pushing the boundaries for quantum supremacy further away, as quantum physicists race to make improvements in their own field. Still, Lidar doesn’t rule out the possibility of a quantum supremacy demonstration in the coming months.

The Google-NASA space agreement has set a timeline for supremacy: another milestone in the race to a scalable quantum computer and one that puts other quantum powers on notice. Whether successful or not the Google-NASA space agreement will establish a new practical standard for measuring quantum supremacy which could be universally adopted. This is important as the transformation of quantum theory into real products with user friendly software will require the development of common standards that are both scientifically rigorous and commercially applicable.  Stay tuned for the results – and the Project Q documentary.

 

 

 

 

Artificial Intelligence

The Making of Ethical Machines


No Comments
BLADE RUNNER 2049Niander Wallace pondering the creation of life altering new technologies in Blade Runner 2049. Image credit: Sony Pictures/Warner Brothers.

Alexander Vipond

Researchers at MIT have undertaken the world’s largest survey on the moral preferences of people to different variations of the trolley problem. The trolley problem’s basic premise is this: a vehicle is about to have an unavoidable accident and the driver must make a choice as to who or what the vehicle hits e.g. swerve right and hit a young man or swerve left and hit two old people?

Edmond Awad and his team collected over 39.6 million decisions from 233 countries through a specially designed mobile game and website. The game and website asked participants to weigh the ethical issues of different versions of the trolley problem according to 9 life indicators (which can be seen in infographic b below). Previously, most studies have relied on single indicators such as a preference for saving many lives over one rather than attempting to look at the complex interrelationships of multiple indicators. From the responses, the researchers were able to discern large scale patterns and trends from 130 of the countries to identify peoples’ key ethical preferences for the preservation of human life.

Moral Machine infographic

Hierarchical clusters of countries based on average marginal causal effect. One hundred and thirty countries with at least 100 respondents were selected. The three colours of the dendrogram branches represent three large clusters—Western, Eastern, and Southern. Country names are coloured according to the Inglehart–Welzel Cultural Map 2010–2014. Image Credit: Awad et al in Nature, ISSN 1476-4687.

They discovered three different ethical worldviews: The Eastern, the Southern and the Western (as displayed in infographic A). These groups agreed on some basic principles and diverged on others. They shared three major preferences. That young people should be spared over others, that many people should be spared over a few and that humans should be spared over other species. These preferences traversed different cultural, economic, political and religious boundaries.

However, as you can see in the radar plots of infographic b, Eastern, Southern and Western views also express sharply different preferences across the spectrum of the nine life indicators. The Western view skews towards saving the young, the many and taking no action at all, giving the choice to chance. The Eastern view skews towards saving the lawful, humans and pedestrians whilst the Southern view prioritises women, the young and high-status individuals.

Machines are on the verge of being programmed to make life-altering choices, a turning point in history. The questions Awad’s team raise over whether universal machine ethics are possible and whether societies can reach consensus over the use of intelligent technologies are a crucial step in discussing what sort of world we want to live in as we undergo the Fourth Industrial Revolution.

While the world is focused on the threat of killer machines on the battlefield, machine decision-making will pose challenges in times of war and peace. This research tests the limits of universal standards as country specific preferences emerge from the complexity of weighing multiple factors. The scalability of new intelligent technologies may be limited by their adaptability to different cultural environments with varying ethical standards. Geo-strategic tensions and ethical dilemmas over who has the power to control these choices, the diversity of datasets used to make technology and the research used to justify life altering choices will affect company, consumer and government.

For example, moving to a different country in the future may mean moving to a set of new technological moral compasses which will have different criteria, levels of access and personalisation, dependent on the rules of the society.

The Moral Machine experiment is only a snapshot in time; a poll of preferences that remains fluid. Ethical standards will require sensible discussion and update periods to reflect changes in the community. Awad notes that the situations presented rely on 100 per cent certainty of the events occurring and 100 per cent certainty of recognising the targets. In the real world there is a much greater level of uncertainty in these processes.

Beyond this lies the extreme technological challenge for engineers and scientists of how to weigh the vast array of preferences with any semblance of granularity. Can your car accurately evaluate someone’s societal status in the 3.2 seconds before a crash? That technology has yet to arrive. However, in some countries the autonomous car might link with the mobile phones of surrounding pedestrians and choose the person with the lowest social credit score by proxy.

As intelligent and networked technologies continue to develop and impact our lives they will increasingly become imbued with formalised versions of the rules that govern our societies. The collective may gain power over the individual. What we have previously left to chance and split-second decision making, we will now expect to be pre-programmed with precision into machines. As Awad’s research shows countries and communities need to start national and regional conversations about what should be delegable to an autonomous machine and how it is operated, before it is decided for them.

 

 

 

Uncategorized

The Politics of Electoral Security


No Comments

Children hack voting machines in the DEF CON voting village. Image via PBS

Dr Aim Sinpeng

“You see I am changing who the winner is in this election” an 8-year-old boy tells me while he is typing away in what looks like some alien language. Then I see the screen. Bob da Builder has suddenly won the Florida state election.

My jaw dropped. This is scary. The boy is only a few years older than my son and he can hack into a US election website to change the winner and the results. Things got scarier as I learned that some of the vulnerabilities in today’s voting machines were identified over a decade ago and are still left unchanged.

“It’s really easy [to hack the election]. I bet the Russians could do it in their sleep… It’s not that secure. I think America should work on its computer security,” says another 11-year-old girl who hacked into the mock voting website in 10 minutes.

The organizers of DefCon’s Voting Machine Village felt terrified but unsurprised by how insecure America’s election infrastructure is.

Some of the vulnerabilities found by participants of the Voting Village were embarrassingly basic. As reported by the Wire, one machine had a root password of “password” and the admin password as “pasta.”

Praise and criticism swirled around the hacking of voting machines at DefCon. Supporters appreciated the lengths to which organizers had gone to expose computer vulnerabilities at the heart of America’s democracy – the voting machines. More than a hundred election officials have attended DefCon to learn, but others have condemned the white-hat hacking. They claim the environment is nothing like the real one on election day. To make matters worse, officials worry that DefCon’s hacking will discourage voters from showing up at the polls.

It’s undeniable however that the actions of DefCon’s Voting Machine Village have had a real impact on people with the power to bring about change. For a time, there was bipartisan support to fix the broken election system. The Secure Election Act was introduced in 2017, requiring election officials to have back-up ballot papers and to conduct a post-election audit. The new bill seemed reasonable, at first, to both Republican and Democrat senators. However it has now been put on ice due to fierce opposition from election officials who bitterly complained about the lack of the funds that would be required to carry out this new mandate. Worse, some early advocates of the election security bill have now backed up – believing the bill will not go far enough in improving the security and integrity of America’s elections. In essence, it’s now dead in the water.

The American public is concerned about the security of their elections. A recent poll by the Pew Research Centre showed that fewer than 50% of Americans feel confident in their election security. Nearly 70% of Americans think it’s likely the Russians and other foreign governments will try to interfere in the mid term elections happening next week.

So there seems to be public support but no political will to change when it comes to the security of American elections. But as the organizers of the Voting Village at DefCon emphasize: money talks. Resources need to be spent to improve electoral security otherwise the Secure Election Act, if given another life, would be an unfunded mandate.