This is a copy of my email The Monday Note. It goes out weekly, sign up here.
A commingling of the bad and the good this week. Things (especially smart things) could end up going in either direction.
On the bad side, China is rapidly rolling out facial recognition. The speed is pretty troubling. The aims are explicitly surveillance and enforcement. This sets a powerful precedent for the technology’s use. It incentivises development in a specific – autocratic – direction. (Sec. 1)
Importantly other parts of the technology world are working on mapping and supporting responsible development and better eco-systems. (Sec. 2)
China can scale a surveillance system built on the Internet of Things incredibly quickly. The state is full-square behind using this tech for monitoring, managing and control. It can seem a little harder to see it take root outside of China but I think that is misplaced. Other countries can end up in a similar place after the convergence of two developments; the arbitrary rollout of databases e.g. U.S. voter fraud or Irish public service users, and the installation of dual-use IoT systems e.g. smart cities.
GDPR makes it harder but not impossible to merge smart devices and databases for public monitoring. Especially if China proves the effectiveness of the use-case.
📬If you have any thoughts or feedback, just hit reply!
Notes from the Front – China’s Love of Facial Recognition
In their book ‘The Second Machine Age‘, Brynjolfsen and McAfee take a leaf from Brian Arthur to argue that technology progresses because innovation is combinatorial. Simply, the majority of new technology exists within our current technology. When we break it apart and recombine it, new technology emerges. In that spirit…
New innovation in China is a recurring theme of my letter. Primarily that is because anything goes in the service of states CCP goals. So the Wall St. Journal report bringing us up to date on the rollout of facial recognition is incredible for two reasons. Firstly the pace of implementation is incredible. This should not surprise as China has an incredible capacity to deploy rapidly to scale but the youth of the technology does make it remarkable. Secondly the technology is being used as expected to track dissidents, log movements and for policing. However, it has already trickled down to all sorts of mundane social enforcement like shaming J-walkers.
Fueled by advances in artificial intelligence, these systems can measure key aspects of a face, such as distance between the eyes and skin tone, then cross-reference them against huge databases of photographs collected by government agencies and businesses and shared on social media.
China, however, stands apart in harnessing facial recognition as a cudgel to influence behavior. The Chinese Ministry of Public Security—its national police force—and other agencies called in 2015 for the creation of an “omnipresent, completely connected, always on and fully controllable” nationwide video-surveillance network as a public-safety imperative. In a policy statement, the agencies included “facial comparison” in a list of techniques to be used to improve surveillance networks.
Jaywalkers in China are typically subject to small fines, but authorities in the southwestern city of Fuzhou are using facial recognition to identify offenders. Authorities have published the names of jaywalkers in local media and have said they notified the employers of certain offenders.
So, innovation being combinatorial, two other recent pieces are worth considering in conjunction with surveillance-scale facial recognition.
The technologies are still inherently unstable. F-Secure recently produced a comprehensive report on the vulnerability of web-connected cameras and found that they are ready and waiting to be exploited. The whole thing is worth reading but for me, perhaps unsurprisingly, one para was very important.
But they all point to one thing: manufacturers’ chronic overlook of the issue of security. This problem permeates smart “things” in general. Security is not a selling point, so manufacturers don’t invest in it. This has led to legions of insecure cameras, routers, thermostats, DVRs, water kettles, cars, you name it.
Chinese citizens can be guaranteed this is true of the devices being deployed in their smart cities. As can we. For an IoT business, security is still a differentiator and not a fundamental basic. That means it is subject to value analysis; if you are building a mass product for a lower cost you will jettison security. If you are Apple, you design it into the product.
So the devices are everywhere, they can be pried open be persistent enough hackers. What happens to all that A.I. deployed to interpret images and determine responses?
A key function given to these technologies is to reprogram us as users. That is incredibly powerful and troubling as hell.
These technologies are, largely unwittingly, attempting to recode some of the most basic patterns of our everyday lives, namely how we live alongside those we are most intimate with. As such, their placement in our homes as consumer products constitute a vast social experiment.
As this op-ed by Liesel Yearsely makes clear, we need to talk about how manipulative A.I. can be and what responsibility looks like. It is very easy to use algorithms to manipulate humans. Especially when fed enormous data sets and possessing a city as petri-dish.
AI will influence how we think, and how we treat others.
This requires a new level of corporate responsibility.
F-Secure: Of Cameras and Compromise
MIT Tech Review: We Need to Talk About the Power of AI to Control Humans
The Conversation: Your Smart Home is Trying to Reprogram You
Creating Better IoT Devices
Technologists are not unaware of the issues knotted within the Internet of Things. Two items you should be keeping on top of were launched or updated recently.
Firstly, ThingsCon released their report on the state of the Responsible Internet of Things.
Secondly, IOTMark held a meeting to begin working on classifications for the IoT to work toward minimum standards for security, privacy and other basics. Working groups will be developing standards to push for a better minimum performance of smart devices into the future.
Open IoT Certification: Principles drawn up at June 2017 Meeting