As facial recognition use grows, so do privacy fears

Above, a facial recognition system for law enforcement during the NVIDIA GPU Technology Conference in November 2017, which showcases artificial intelligence, deep learning, virtual reality and autonomous machines. (AFP)
Updated 08 July 2018
0

As facial recognition use grows, so do privacy fears

  • Facial recognition is playing an increasing role in law enforcement, border security and other purposes in the US and around the world
  • While more accurate facial recognition is generally welcomed, civil liberties groups say specific policy safeguards should be in place

WASHINGTON: The unique features of your face can allow you to unlock your new iPhone, access your bank account or even “smile to pay” for some goods and services.
The same technology, using algorithms generated by a facial scan, can allow law enforcement to find a wanted person in a crowd or match the image of someone in police custody to a database of known offenders.
Facial recognition came into play last month when a suspect arrested for a shooting at a newsroom in Annapolis, Maryland, refused to cooperate with police and could not immediately be identified using fingerprints.
“We would have been much longer in identifying him and being able to push forward in the investigation without that system,” said Anne Arundel County police chief Timothy Altomare.
Facial recognition is playing an increasing role in law enforcement, border security and other purposes in the US and around the world.
While most observers acknowledge the merits of some uses of this biometric identification, the technology evokes fears of a “Big Brother” surveillance state.
Heightening those concerns are studies showing facial recognition may not always be accurate, especially for people of color.
A 2016 Georgetown University study found that one in two American adults, or 117 million people, are in facial recognition databases with few rules on how these systems may be accessed.
A growing fear for civil liberties activists is that law enforcement will deploy facial recognition in “real time” through drones, body cameras and dash cams.
“The real concern is police on patrol identifying law-abiding Americans at will with body cameras,” said Matthew Feeney, specialist in emerging technologies at the Cato Institute, a libertarian think tank.
“This technology is of course improving but it’s not as accurate as science fiction films would make you think.”
China is at the forefront of facial recognition, using the technology to fine traffic violators and “shame” jaywalkers, with at least one arrest of a criminal suspect.
Clare Garvie, lead author of the 2016 Georgetown study, said that in the past two years, “facial recognition has been deployed in a more widespread and aggressive manner” in the US, including for border security and at least one international airport.
News that Amazon had begun deploying its Rekognition software to police departments sparked a wave of protests from employees and activists calling on the tech giant to stay away from law enforcement applications.
Amazon is one of dozens of tech firms involved in facial recognition. Microsoft for example uses facial recognition for US border security, and the US state of Maryland uses technology from German-based Cognitec and Japanese tech firm NEC.
Amazon maintains that it does not conduct surveillance or provide any data to law enforcement, but simply enables them to match images to those in its databases.
The tech giant also claims its facial recognition system can help reunite lost or abducted children with their families and stem human trafficking.
Nonetheless, some say facial recognition should not be deployed by law enforcement because of the potential for errors and abuse.
That was an argument made by Brian Brackeen, founder and the chief executive officer of the facial recognition software developer Kairos.
“As the black chief executive of a software company developing facial recognition services, I have a personal connection to the technology, both culturally and socially,” Brackeen said in a blog post on TechCrunch.
“Facial recognition-powered government surveillance is an extraordinary invasion of the privacy of all citizens — and a slippery slope to losing control of our identities altogether.”
The Georgetown study found facial recognition algorithms were five to 10 percent less accurate on African Americans than Caucasians.
Microsoft announced last month it had made significant improvements for facial recognition “across skin tones” and genders.
IBM meanwhile said it was launching a large-scale study “to improve the understanding of bias in facial analysis.”
While more accurate facial recognition is generally welcomed, civil liberties groups say specific policy safeguards should be in place.
In 2015, several consumer groups dropped out of a government-private initiative to develop standards for facial recognition use, claiming the process was unlikely to develop sufficient privacy protections.
Cato’s Feeney said a meaningful move would be to “purge these databases of anyone who isn’t currently incarcerated or wanted for violent crime.”
Jennifer Lynch, an attorney with the Electronic Frontier Foundation, said that the implications for police surveillance are significant.
“An inaccurate system will implicate people for crimes they did not commit. And it will shift the burden onto defendants to show they are not who the system says they are,” Lynch said in a report earlier this year.
Lynch said there are unique risks of breach or misuse of this data, because “we can’t change our faces.”
Evan Selinger, a philosophy professor at the Rochester Institute of Technology, says facial recognition is too dangerous for law enforcement.
“It’s an ideal tool for oppressive surveillance,” Selinger said in a blog post.
“It poses such a severe threat in the hands of law enforcement that the problem cannot be contained by imposing procedural safeguards.”


Can data save the world?: Experts discuss how statistics can help to solve some of our biggest challenges

Updated 11 November 2018
0

Can data save the world?: Experts discuss how statistics can help to solve some of our biggest challenges

  • Connections between the data community and political decision-makers are critical to realize the transformative power of data
  • Figures improve for middle-income countries where knowledge is more widespread and where close to 92 percent of people are identified

DUBAI: The idea of almost 2,000 statisticians meeting to talk about data, as they did at the recent UN World Data Forum in Dubai, is enough to make anyone’s eyes glaze over, but it would be foolish to ignore the important conversations they are having about the power of data to help solve some of the world’s biggest challenges, from poverty to global warming.
For instance, data from satellite imagery and radars can tell us about glaciers melting, deforestation and the state of algae in our oceans.
The potential of data to help with sustainable development was discussed in Dubai last month, along with how to ensure all individuals are accounted for in data collection and how to leverage new technologies. The event was the first in the region, following its first meeting in Cape Town in 2017.
Although the forum is only in its second year, progress has been made with the UN leading work globally through working groups on developing tools, governing systems and the principles that deal with issues of open data and data privacy, including the use of private-sector data.
“We live during a time of unprecedented challenge, but equally unprecedented and massive opportunity,” said Amina Mohammed, deputy secretary-general of the UN. “Our blueprint for addressing these challenges and seizing the opportunities is the 2030 Agenda for sustainable development. But to achieve the 17 Sustainable Goals (SDGs), we will need more and better data. With accurate, representative, inclusive and disaggregated data, we can understand the challenges we face and identify the most appropriate solutions for sustainable development.”
The SDGs were adopted by the UN to end poverty and protect the planet, including areas such as climate change, economic inequality, innovation, sustainable consumption, peace and justice.
The forum looked at how data can play a crucial role in saving and improving lives, whether in disaster preparedness and early warning systems, providing job opportunities for students or educating women about laws protecting them against discrimination. “It can strengthen trust in public institutions and unveil new opportunities,” Mohammed said. “But while it is clear that the data revolution is having an enormous impact, it has not benefited everyone equally.”
Since 1970, natural disasters have affected the lives of more than 460 million people in Africa, many of who could have been saved with better data and forecasting, according to Mohammed. In more than two thirds of countries, there is also a lack of gender disaggregated data on violence against women, which would allow experts to uncover patterns, and consequently, tackle the issue more efficiently.
All these issues work toward achieving the UN’s 2030 Agenda for Sustainable Development. “We are nearing the 2030 Agenda and the SDGs, integrating all three dimensions of sustainability — economic, social and environmental — that will guide international development efforts and national policy through 2030,” said Liu Zhenmin, under-secretary-general for the UN Department of Economic and Social Affairs. “To do so, it is essential to have relevant, timely, open and disaggregated data, which requires that all communities represented today fulfill a critical role and find ways to work across different domains and create partnership and synergies.”
Three years into the 2030 Agenda, Zhenmin said that national data can be used to help implement and monitor it.
“The unprecedented number of new initiatives and approaches for the improvement of data production and utilization raises the importance of data and statistics for the full implementation of the 2030 Agenda,” he said.
According to Mahmoud Mohieldin, senior vice president for the 2030 Development Agenda at the World Bank, “leaving no one behind” is an aspirational objective that must be translated into action by finding out how many people have not been accounted for. “The figures are very worrying,” he said. “For the low-income countries, we have no clue about services and support to around 40 percent of the population because they don’t have any kind of identification. As far as the official records, they don’t exist.”
The number is 6 percent higher for women in those countries, creating significant discrepancy in the field. Figures improve for middle-income countries where knowledge is more widespread and where close to 92 percent of people are identified. In upper middle-income countries, the number reaches almost 97 percent.
“Technology today is really helpful, and I’ve seen great transformation in identification when it has good policy and leadership,” Mohieldin said. “For the new digital economy, there is no way for the public to get access to services or to be part of the new economy without identification, and that identification needs to be electronic, secured and supported by systems. It’s the new DNA – you need better data systems, secured networks and artificial intelligence mastering the new codes and languages of the future.”
Being able to analyze an increasing amount of data is becoming a race against time. “When we look at the SDGs, it’s a challenge and if we don’t have systems to read this data, crunch it and give us advice in real time, we are losing this race,” said Omar Al-Olama, UAE Minister for Artificial Intelligence.
“The second challenge is that we need to be informed on a real-time basis, second by second, and deploying these systems in a way that the data actually (pools) directly into the AI algorithm or system would allow us to take much more informed decisions. We can leverage technology for us to take much more insightful decisions and to achieve the SDGs in the time frame set.”
He spoke of data shaping the future of our planet. “When it comes to data and new platforms, no one has it right and we’re all experimenting together,” Al-Olama said. “We also need to increase data and technology literacy across our companies and governments. People need to understand why it’s important for us to ride this wave to control systems in the future.”
There are currently 350,000 organizations worldwide collecting data for their areas of interest. But with the rise and inevitable evolution of cloud computing, opportunities have emerged for information-sharing and bringing data together from various systems. “We’re already starting to see evidence of that, not just for professionals, but also appropriate access for the public and interested parties,” said Clint Brown, director of product engineering at ESRI. “Systems are coming alive, computing power is becoming available and these cloud systems can do some amazing things.”
He gave the example of the entire Landsat imagery collection going back to 1970s, which is now available online for processing. “They’re huge datasets so we should think as a priority how do we work together, and make data and open up access in appropriate ways,” he added. “This opportunity for many people to participate is upon us.”
As the world prepares to take in an abundance of data in the near future, improved information systems, infrastructure and support in analysis will be needed. “Data and the statistical community are placed at the heart of driving the SDGs Agenda, which is a big responsibility that’s been given to all of us,” said Harpinder Collacott, executive director of Development Initiatives, an independent international organization that focuses on the role of data in driving poverty eradication and sustainable development. “What we do with that responsibility is really going to be critical in the implementation of the global goals.”
She said the voice of policy-makers, from national to sub-national, was particularly relevant in the field. “I’m not sure how much of it we will hear at the forum,” she said. “But it’s their needs that have to be driving the investments around the data that we produce so that they can make the right decisions and put in place the right policies and approaches to really implement the SDG agenda.”
Connections between the data community and political decision-makers are critical to realize the transformative power of data. “Data has to move beyond just improving data systems to improving people’s lives,” she added. “So we need to demonstrate that action pretty quickly, which is a challenge because we’re nearly a quarter of the way through the SDG agenda already and we need to think about how we are using the data that is already available — and we have huge amount — so we can start to spearhead some action and improve data systems, governance and interoperability as we go forward.”