The NYT reported that smart assistants can be (re-)programmed by ultrasonic signals, which are inaudible to their owners. Image Source: NYT / Lynn Scurfield.
There have been several headlines about AI and robots in recent weeks. All of them involve a gap between what we want to see in technology (and in ourselves) and what we don't want to see.
Introducing SpotMini (23 June 2016). Video Source: Youtube.
Boston Dynamics CEO on being acquired and selling the SpotMini | TC Sessions Robotics 2018 (11 May 2018). Video Source: Youtube.
Silicon Valley spins technology as helpful and progressive. While this is partly true, there is obviously a potentially abusive side to innovation. These headlines embody that dichotomy:
- Facebook has delayed the announcements on its line of home AI products in light of negative publicity around Mark Zuckerberg's interrogation by Congress. These new AI products undoubtedly include Facebook technology which can scan your thoughts inside your head, convert them into text, and publish them in your newsfeed. This tech was announced by Regina Dugan, VP of Engineering at Building 8 at the company's F8 conference in April 2017. The mind reader was originally developed at Stanford to help paralyzed people type using their minds. Dugan further stated: "We also described a system that may one day allow you to hear through your skin." Now Facebook is seeking general application of the tool for its users. But don't worry: "Facebook is only translating the thoughts you want to share."
- Boston Dynamics is encouraging people to buy its killer robot dog, SpotMini, to load their dishwashers or help out at the office. SpotMini goes on sale in 2019.
- The New York Times reported that Amazon Alexa, Google Assistant, and Apple's Siri can receive ultrasound commands that are inaudible to human ears, meaning the thing that controls your lights, locks, smart fridge, alarm system, and other areas of a residence related to life and security can be hijacked by someone else around the property or online. Also, incoming music can also be embedded with commands to these devices which cannot be heard by humans. This paves the way for ultrasonic attacks on civilians, by criminal, state, or non-state actors. NYT: "A group of students from University of California, Berkeley, and Georgetown University showed in 2016 that they could hide commands in white noise played over loudspeakers and through YouTube videos to get smart devices to turn on airplane mode or open a website. This month, some of those Berkeley researchers published a research paper that went further, saying they could embed commands directly into recordings of music or spoken text. So while a human listener hears someone talking or an orchestra playing, Amazon’s Echo speaker might hear an instruction to add something to your shopping list."
- Google's disturbing vision of the future was revealed in a leaked video, The Selfish Ledger: "Leaked video reveals a Black Mirror-style future in which technology could be used to control the behaviour of entire populations" You can watch the video below, or here.
- After launching a protest petition with nearly 4,000 signatures, "about a dozen" of Google's employees resigned over the company's involvement in lending its AI to develop drone warfare contracts under Project Maven. Google is also competing to win another US government cloud contract, the Joint Enterprise Defense Infrastructure (JEDI). These are the first mass resignations which have ever occurred in a protest against the company's policies. Those who departed submitted statements, including: "It’s not like Google is this little machine-learning startup that’s trying to find clients in different industries. ... It just seems like it makes sense for Google and Google’s reputation to stay out of that." And: "I wasn’t happy just voicing my concerns internally. The strongest possible statement I could take against this was to leave."
- The Tech Workers Coalition also launched a petition against Google, Microsoft, IBM, and Amazon, asking that the companies not lend their AI and other tech to Pentagon contracts and military applications. The petition "asks for industry-wide ethical standards for the use of artificial intelligence."
- Via the International Committee for Robot Arms Control, 90+ academics who work in the field of AI also penned an open letter asking that Google's top executives not put the company's tech to military use: "Should Google decide to use global internet users’ personal data for military purposes, it would violate the public trust that is fundamental to its business by putting its users’ lives and human rights in jeopardy. The responsibilities of global companies like Google must be commensurate with the transnational makeup of their users. The DoD contracts under consideration by Google, and similar contracts already in place at Microsoft and Amazon, signal a dangerous alliance between the private tech industry, currently in possession of vast quantities of sensitive personal data collected from people across the globe, and one country’s military. They also signal a failure to engage with global civil society and diplomatic institutions that have already highlighted the ethical stakes of these technologies." One of the petition's authors, Lucy Suchman, raised the question of US military priorities versus Google's obligation to protect its global user base, and the trust that base has placed in the company not to use their personal data in applications that could be militarily aimed at them.
- In the same week, Google unveiled AI Assistant voices at its I/O developer expo; the voices, based on voice actor recordings, are indistinguishable from real humans. The AI tool Google Duplex allows the Assistant to telephone businesses on your behalf and make appointments. The company presented a convincing demo, which some critics claim is enhanced or faked.
- At the same expo, Google promised its AI will increase censorship of unapproved news commentators to keep us safe online. The AI will screen for fake news and make sure we are not exposed to any 'dangerous ideas.' The AI presents the 5 top headlines to you, based on your preferences. Google News will offer fact checks, opinions, and analyses. An unfiltered view of events will be available, but only from “trusted news sources.”
- China has developed a social credit index. The AI monitors all your activities and ranks your value as a citizen. With a low score (acquired by doing things like criticizing the government, driving poorly, or smoking cigarettes in non-smoking zones), you can be blocked from buying property, traveling, patronizing particular businesses, and staying at certain hotels. Your Internet access can be throttled. Children whose parents have low scores can be refused admission to certain schools. You can also be publicly exposed as a bad citizen and blocked from entering certain professions. Those with high scores can get deals at banks and supermarkets. See reports: here, here, here, and here.
- Chinese AI is being used in schools to monitor children and teens with facial recognition to inform the teacher about which students are not paying attention. Students are frightened by the changes: "One child in the class told Hangzhou, 'Previously when I had classes that I didn't like very much, I would be lazy and maybe take a nap on the desk or flick through other textbooks. But I don't dare be distracted since the cameras were installed in the classrooms. It's like a pair of mystery eyes are constantly watching me.'"
- Over a dozen Chinese companies have embedded AI sensors in workers' helmets: "Some construction, manufacturing and transport companies in China require workers to wear helmets fitted with brain sensors that use AI technology to monitor their emotions. Spokespeople for the scheme say that the system saves money by helping workers and training them better, but critics worry about the control that companies have over employees. The devices fit onto the helmets of workers which then measures the brain activity of the wearer, these waves can be analyzed to look for changes in emotion. Spikes and dips in the brainwaves can be read as changes in emotions and feelings associated with panic, fatigue, and sadness." Interesting Engineering reports: "One official who runs the 'emotional surveillance' program at the State Grid Zhejiang Electric Power told Chinese media that the devices are essential to running an efficient business. The company’s profits have reportedly grown by $315 million since the company began using the devices in 2014." South China Morning Post reports: "‘Forget the Facebook leak’: China is mining data directly from workers’ brains on an industrial scale."
The leaked video that reveals Google's 2016 vision of the future and published by The Verge: Google’s The Selfish Ledger (leaked internal video) (17 May 2018). Video Source: Youtube.
What else can surveillance cameras do in classroom other than exam supervision? High school in #Hangzhou uses camera to identify students facial expression for class performance analysis and improvement pic.twitter.com/bXolAE7Ev8— People's Daily,China (@PDChina) May 16, 2018
The Paranoid's Worst Case Scenario
People have spent years on Google, searching for what they want to know. Meanwhile, Google has collected that data and built up a model of collective human thoughts. On the basis of this model, the company has built a mirror mind, an outer brain, an AI version of the collective consciousness.
In my previous post on this subject, I remarked that AI will be a mirror of ourselves. It will reflect our conscious and unconscious impulses. That means there will be two sides to this story. The shiny narrative of progress, and the shadowy unconscious.
On the one hand, this technology is being applied in impressive AI services which heal diseases, help people solve problems, and smoothly negotiate 21st century life. On the other hand, all that data on GPS tracking, collective needs, desires, and group behaviours makes a perfect template for drones to track down and kill targeted insurgents in war zones. For its part, Google denies that its AI contributions to the US drone program will be applied to killing people.
A paranoid would say that drone warfare could come home and be used on civilian populations. Imagine the Chinese social credit score universally applied in tandem with western AI censorship, which weeds out people who have the wrong economic credit scores, the wrong opinions, and the wrong online friends.
In March 2018, the New York Times reported on a Consumer Watchdog study that scrutinized Amazon's and Google's smart assistant patents. Despite reassurances that these companies respect users' privacy, the study (here) found that smart assistants are programmed to listen passively, detect the emotions of people in their vicinity, and to record users' bathing habits. The machines also assess whether or not children in the house are engaging in mischief and the machines record and report those findings! The AI identifies people in the household and builds profiles on each one.
Superficially, there are marketing applications for these data. But really, the game plan is bigger and by-passes the whole legal and judicial system. If there is one thing we should guard against, it is passivity and complacency. At present, Julian Assange has been cut off from Internet access, visitors, and other communications for over 50 days and no one bats an eyelash. As an individual, he has courted danger, but his situation also sets a precedent.
On the one hand, this technology is being applied in impressive AI services which heal diseases, help people solve problems, and smoothly negotiate 21st century life. On the other hand, all that data on GPS tracking, collective needs, desires, and group behaviours makes a perfect template for drones to track down and kill targeted insurgents in war zones. For its part, Google denies that its AI contributions to the US drone program will be applied to killing people.
A paranoid would say that drone warfare could come home and be used on civilian populations. Imagine the Chinese social credit score universally applied in tandem with western AI censorship, which weeds out people who have the wrong economic credit scores, the wrong opinions, and the wrong online friends.
In March 2018, the New York Times reported on a Consumer Watchdog study that scrutinized Amazon's and Google's smart assistant patents. Despite reassurances that these companies respect users' privacy, the study (here) found that smart assistants are programmed to listen passively, detect the emotions of people in their vicinity, and to record users' bathing habits. The machines also assess whether or not children in the house are engaging in mischief and the machines record and report those findings! The AI identifies people in the household and builds profiles on each one.
Superficially, there are marketing applications for these data. But really, the game plan is bigger and by-passes the whole legal and judicial system. If there is one thing we should guard against, it is passivity and complacency. At present, Julian Assange has been cut off from Internet access, visitors, and other communications for over 50 days and no one bats an eyelash. As an individual, he has courted danger, but his situation also sets a precedent.
— WikiLeaks (@wikileaks) May 12, 2018
Your value as a citizen is already being assessed, because your laptop, mobile phone and television are hot microphones which record your conversations. Depending on the device and jurisdiction, your computer is recording your keystrokes and the government is retaining your entire Internet browsing history. Connect that to AI which controls: (1) your home's lights; (2) the locks on your residence and inside your home; (3) your electronic-enabled car or driverless vehicle; (4) your refrigerator, stove and oven; (5) your grocery order; (6) your telephone, television, and computer; (7) your bank and credit access.
Imagine a nightmarish future in which your social credit score drops. Suddenly, your Internet connection and social media accounts are closed down and erased, and your online followers can no longer find you. Or someone else - whether a real person or an AI smart assistant - takes over your social media accounts and begins sending e-mails and posting opinions that aren't yours.
Or perhaps you cross an invisible line, and when you do, your Internet search results become increasingly misleading. Your television starts showing you programs and newscasts which no one else has seen, increasing paranoia and isolation. Your residence locks you in or out, or confines you to certain rooms in your home. The home lights maintain darkness or constant brightness; or your smart assistant begins blaring music constantly. These are modern torture techniques; recent conflicts mean that the US government has hard data on the effects of enforced sleeplessness.
Or maybe your grocery delivery stops and/or your smart fridge refuses to open for you. Your smart assistant, which sounds just like a real person, calls your employer, friends, or family 'on your behalf.' Your smart car locks you inside and drives you off a bridge or into a wall. Or it simply drives into you. Or perhaps it refuses to operate in case you need to leave town on short notice. Or it gets hacked. A malevolent totalitarian state could employ these methods to commandeer your tech as easily as any future hacking group, criminal gang, or state-employed, corporate-hired, or freelance mercenaries.
Real Red Flags
Actor Anton Yelchin Dies From a Car Crash at His Own House (20 June 2016). Video Source: Youtube.
In 2017, the WikiLeaks Vault 7 release confirmed that smart cars can be remotely controlled. The Hacker News:
That is before we get to elevators, escalators, street cars, subways, trains, and airplanes."One leaked document suggested that the CIA was even looking for tools to remotely control smart cars and trucks, allowing the agency to cause 'accidents' which would effectively be 'nearly undetectable assassinations.'"
Michael Hastings Car Accident Still Makes No Sense (5 August 2013). Video Source: Youtube.
What Really Killed Michael Hastings? (Just asking...) (21 August 2013). Video Source: Youtube.
Rolling Stone journalist Michael Hastings was killed on 18 June 2013 in a mysterious car accident after embarrassing US General Stanley McChrystal for his actions in Afghanistan. Hastings had contacted a WikiLeaks lawyer for help hours before he died. He was driving a Mercedes C250. The Hastings case became a favourite of conspiracy theorists and alt-news Youtubers. When he died, Hastings had just finished an exposé on CIA director, John O. Brennan. In January 2014, Michael's brother Jonathan Hastings gave an interview in which he ruled out foul play in his brother's demise.
At the time of his death, Michael Hastings was also developing a profile piece on the American journalist, Barrett Brown, who is now linked to the new Internet Party (see my related post on Brown and the party, here). Brown's problems started when he began investigating the hacker group, Anonymous:
Brown was imprisoned for sharing a link to an enormous trove of hacked information from Stratfor, regarding American government surveillance of US citizens. The leak additionally revealed a lot of insider information on global affairs, the corruption of professional journalists, and corporate misdeeds. In relation to the Stratfor leak of over five million internal e-mails, Brown was imprisoned from 2012 to 2016 and again briefly in 2017. The original hacker in the case was Jeremy Hammond. You can read WikiLeaks' statement on the Stratfor material here and you can search the files here."In early 2010, journalist and satirist Barrett Brown was working on a book on political pundits, when the hacktivist collective Anonymous caught his attention. He soon began writing about its activities and potential. In a defense of the group’s anti-censorship operations in Australia published on February 10, Brown declared, 'I am now certain that this phenomenon is among the most important and under-reported social developments to have occurred in decades, and that the development in question promises to threaten the institution of the nation-state and perhaps even someday replace it as the world’s most fundamental and relevant method of human organization.'"
My next post on AI will consider the implications of these developments for global politics.
Crashes of Convenience: Michael Hastings (5 July 2013). Video Source: Youtube.
No comments:
Post a Comment