Feeds:
Posts
Comments

Posts Tagged ‘the algorithm is not your friend’

“The phone rings and an eerie voice mimicking a woman says: “Your shift is about to start, please log in and make sure you are in your designated starting area.”

This automatic reminder to promptly start the shift was, for me, the most tangible reminder that couriers working for Foodora and similar “platform capitalism” companies operate via the control of algorithms. While getting a call from a robot urging to start your shift is uncanny, the algorithms control the couriers’ work in more subtle, and more important, ways.

Dispatching, namely the assignment of orders to couriers is done automatically at Foodora. Through the application that the couriers use to receive orders, the algorithm tracks the couriers’ location, average speed, how quickly the courier delivers the food and how much time they spend at the customer. Based on an unknown weighing of these factors, the algorithm assigns a specific order to a given courier.

The dispatcher, who distributes orders, plays probably the single most important part in a courier’s job. The courier plans their own routes, but the dispatcher gives the orders, sets the pace and provides the information the courier needs to do their job. No matter how fast a courier rides or how well they navigate the city, if dispatching does not work, nothing works. Conversely, when dispatchers and couriers work well together and communicate with each other, they deliver orders quickly and efficiently. When the dispatcher is a courier themselves, this cooperation usually works best, because the dispatcher knows what can be expected from a cyclist, how the weather, the load and distance affects the courier.

Foodora has human dispatchers, who oversee couriers in a given city. However, in the working process designed by Foodora, human dispatchers ideally don’t interact with couriers, who should get their orders automatically. Presumably as a cost-saving measure, Foodora centralized its dispatching to Berlin and the dispatchers overseeing say Helsinki know nothing about the city. Thus the dispatchers are not able to help couriers in problem cases and sometimes the results are just plain bizarre, for example when by mistake an order has registered to a restaurant that is in fact closed and the courier tries to tell disbelieving dispatchers in Berlin that the cannot pick it up, because… well, the restaurant is closed.

The biggest problem however is one of transparency. The provisions paid for the order form a substantial part of the couriers’ income at Foodora, and because of this, those who get more orders earn more. The courier however does not know how and why the algorithm distributes the orders to one courier instead of another. Apparently, the algorithm distributes orders to couriers it deems “effective”. I have seen a situation when a fast courier came exhausted with less than ten minutes of their shift remaining to the office where couriers, who had just started their shifts sat waiting for orders. Then a new order came and algorithm assigned it to the fast courier. Why, nobody knows, but in Foodora’s automatic system once an order is assigned it cannot be changed.

Similarly, the algorithm classes Foodora’s couriers into four “batches”, or groups, based on their performance (as judged by the algorithm). Shifts are made available in steps to these batches so that the first batch, with the “best” couriers, get first pick from all the shifts, then the next and those in the last batch pick any shifts that might be left. How a courier gets shifts obviously directly affects their income. If one can do only a limited amount of hours, one also earns less. Along with this direct effect, how much and how well one works affects also one’s position in the “batch” and the possibility to get shifts in the future.

In short, the algorithms directly control the couriers’ work and their income, but in ways the courier can only guess. Even if the courier was adept in reading the code and reverse-engineering the applications, the systems that manage them are proprietary and not made known to the courier.”

– Tuomas Tammisto,
“When Mr. Robot is Your Boss: Working under algorithms.” The Transnational Courier Federation (#4.2)

When Mr. Robot is Your Boss: Working under algorithms

Read Full Post »

“The rise of ride-sharing services has increased traffic deaths by 2% to 3% in the US since 2011, equivalent to as many as 1,100 mortalities a year, according to a new study from the University of Chicago and Rice University.

How it was calculated: Researchers took statistics from the National Highway Traffic Safety Administration and compared them with the dates Uber or Lyft launched in a specific city. Then they checked accident rates in those cities relative to vehicle miles traveled. That rate shot up in San Francisco after Uber launched in 2010, a phenomenon that was replicated in other cities.

Deadheading: The increase in congestion is partly because drivers spend 40% to 60% of their time searching for passengers, a practice known as “deadheading.” On average, drivers in New York City traveled 2.8 miles between fares.

Before ride-sharing: Traffic deaths fell to their lowest number just before Uber launched in San Francisco. In 2010 there were 32,885 fatal car accidents nationwide, the lowest number since 1949. This decline halted and then reversed after the introduction of ride-sharing in US cities. However, it “may be too soon to tell whether the effect we document is a short-term adjustment or a longer-term pattern,” the researchers said.

Piling up problems: The study adds to a growing body of research on ride-sharing companies. Recent studies have found they increase congestion and cut the use of public transport. Cities are starting to respond to harms, perceived or otherwise. New York’s city council introduced a cap on ride-sharing in August, for example.”

– Charlotte Jee, “Uber and Lyft are behind a sharp rise in US traffic deaths.” The Download, MIT Technology Review. October 25, 2018.

Read Full Post »

“The research tacks against the idea that younger people who are extremely online (or “digitally savvy,” in Pew’s terms) might be more exposed and/or more susceptible to misinformation. But the real correlation with poor performance is exposure to television news, which has fallen off among young people but remains very high among older people. This shouldn’t be surprising if we consider the evolution of American media over the past 60 years. Someone born in 1958, now 60, witnessed two revolutions in media before the internet: talk radio and 24-hour cable news. Both blended facts and opinions in new and unprecedented ways, and they matured with the cohort of Americans who are now over the age of 50.

In 1987, the Reagan administration repealed the FCC’s Fairness Doctrine. That paved the way for the rise of right-wing talk radio, brilliantly chronicled by David Foster Wallace for this magazine. Describing a talk-radio host, John Ziegler, Wallace noted that it was not his job “to be responsible, or nuanced, or to think about whether his on-air comments are productive or dangerous, or cogent, or even defensible.” He has only to be “stimulating.”

Earnest critiques of the facts and opinions that Ziegler put into the world as if he were a journalist made no sense. “Maybe it’s better to say that he is part of a peculiar, modern, and very popular type of news industry, one that manages to enjoy the authority and influence of journalism without the stodgy constraints of fairness, objectivity, and responsibility that make trying to tell the truth such a drag for everyone involved,” Wallace concluded.

Sound familiar to anyone? While talk radio caters to all tastes, the medium developed to serve an audience Pew described in 2004 as “a distinct group; it is mostly male, middle-aged, well-educated and conservative.” That cohort is now over 50, and its members spent decades listening to radio hosts stimulate by mixing facts and opinions in whatever proportion was necessary to keep listeners from turning the dial.”

– 

ALEXIS C. MADRIGAL, “Older People Are Worse Than Young People at Telling Fact From Opinion.The Atlantic. October 23, 2018.

Read Full Post »

“At the center of the criticism was data mining company Palantir, which designed the Investigative Case Management system. The ICM is a critical component of ICE’s deportation operations—it integrates a vast ecosystem of public and private data to track down immigrants and, in many cases, deport them.

Little is known about how the software actually works or how extensively ICE uses it. But within the first nine months of the Trump administration, ICE arrests increased 42% compared with the same period in the previous year. According to civil rights and immigration activists, ICM is fueling the mass surveillance and targeting of immigrants at an unprecedented scale.

Now a new investigation, published today, sheds more light on the web of tech companies involved in supporting ICE and its parent agency, the Department of Homeland Security.

The report, commissioned by activist organizations Mijente, the National Immigration Project, and the Immigrant Defense Project, found that Amazon has played as central a role as Palantir in providing the backbone infrastructure for many of ICE’s, and DHS’s, key programs. Amazon has also enjoyed a cozy relationship with the federal government that has helped it secure an outsize number of government contracts.

“What we’re starting to see more and more is that technology and technology contracts form a huge part of ICE’s budget and are also one of their critical tools for how they’re conducting enforcement on the ground,” says Jacinta Gonzalez, the field director at Mijente.

POWERED BY AMAZON

In 2017, an Intercept investigation found that ICM pulled together data from an array of federal and private law enforcement entities to create detailed profiles that were then used to track immigrants. That data could include a person’s immigration history, family relationships, personal connections, addresses, phone records, biometric traits, and other information.

All of that data and the algorithms powering ICM are now being migrated to Amazon Web Services (AWS) in their entirety; Palantir pays Amazon approximately $600,000 a month for the use of its servers, according to the report’s authors.

Though the money doesn’t flow directly from ICE to Amazon, the tech giant had the right incentives in place for Palantir to choose AWS. In order for Palantir to secure its contract with the government, ICM had to be hosted on a federally authorized cloud service. An online government database shows that Amazon holds the largest share, 22%, of federal authorizations under the FedRAMP program, which verifies that cloud providers have the necessary security requirements to process, store, and transmit government data. More important, Amazon holds 62% of the highest-level authorizations, usually needed to handle data for law enforcement systems.

In a sense, Amazon was merely capitalizing on a trend. In 2010, the US government established a “cloud first” policy and began moving its agencies’ data and computing resources to the cloud. That was cemented in 2014 with the passage of the Federal Information Technology Acquisition Reform Act (FITARA). As the legislation was moving through Congress in January of that year, Amazon, Microsoft, and EMC (since acquired by Dell) formed a lobbying group called the Cloud Computing Caucus Advisory Group to help push it through. The three companies’ PACs also contributed over $250,000 in direct campaign contributions to the two members of Congress sponsoring the act, the Mijente report found.

Additionally, DHS was among the earliest agencies to adopt Amazon cloud services under Mark Schwartz, chief information officer at the US Citizenship and Immigration Service (USCIS). In 2017, after facilitating a major migration of one of DHS’s sub-agencies to AWS, Schwartz left the agency to become the enterprise strategist at that company. AWS did not respond to MIT Technology Review’s request to speak with Schwartz about his relationship with the company during his time in government.

In addition to powering ICM, AWS hosts several of DHS’s other major immigration-related databases and operations, including all the core data systems for USCIS and biometric data for 230 million individuals, including fingerprints, face records, and iris scans, which are playing a growing role in immigration enforcement around the country.

There is no publicly available data on how much Amazon profits from these contracts, but DHS’s complete IT portfolio totals $6.8 billion, which accounts for close to 10% of the agency’s projected spending in fiscal year 2019. An AWS spokesperson had no comment when presented with details of the new report.

Amazon is now also bidding for a $10 billion contract with the Department of Defense to modernize the agency’s computing infrastructure and integrate all US military operations into a single platform. Because of the company’s existing dominance among the government’s cloud providers, it is widely expected to win the contract.”

– Karen Hao, “Amazon is the invisible backbone behind ICE’s immigration crackdown.” Intelligent Machines, MIT Technology Review. October 22, 2018.

Read Full Post »

“As it happens, tech – or surveillance capitalism – has disrupted the private investigation business as much as it’s ripped through journalism, the taxi business, war making, and so many other private and public parts of our world. And it’s not only celebrities and presidential candidates whose privacy hackers have burned through. Israeli spyware can steal the contacts off your phone just as LinkedIn did to market itself to your friends. Google, the Associated Press reported recently, archives your location even when you’ve turned off your phone. Huge online database brokers like Tracers, TLO, and IRBsearch that law enforcement and private eyes like me use can trace your address, phone numbers, email addresses, social media accounts, family members, neighbors, credit reports, the property you own, foreclosures or bankruptcies you’ve experienced, court judgments or liens against you, and criminal records you may have rolled up over the years.

Ten years ago, to subscribe to one of these databases, I had to show proof that I was indeed a licensed investigator and pass an on-site investigation to ensure that any data I downloaded would be protected. I was required to have a surveillance camera and burglar alarm on the building where my office was located, as well as a dead bolt on my office door, a locked filing cabinet, and double passwords to get into my computer. Now, most database brokers just require a PI or attorney license and you can sign right up online. Government records – federal and state, civil and criminal – are also increasingly online for anyone to access.

The authoritarian snoops of the last century would have drooled over the surveillance uses of the smartphones that most of us now carry. Smartphones have, in fact, become one of the primo law enforcement tools other than the Internet. “Find my iPhone” can even find a dead body – if, that is, the victim left her iPhone on while being murdered. And don’t get me started on the proliferation of surveillance cameras in our world.

Take me. I had a classic case that shows just how traceable we all now are. There was a dead body, a possible murder victim, but no direct evidence: no witnesses, no DNA, no fingerprints, and no murder weapon found. In San Francisco’s East Bay, however, as in most big American cities, there are so many surveillance cameras mounted on mom-and-pop stores, people’s houses, bars, cafes, hospitals, toll bridges, tunnels, even in parks, that the police can collect enough video, block by block, to effectively map a suspect driving around Oakland for hours before hitting the freeway and heading out to dump a body, just as the defendant in my case did.

Once upon a time, cops and dirty private eyes would have had to attach trackers to the undercarriages of cars to follow them electronically. No longer. The particular suspect I have in mind drove his victim’s car across a bridge, where cameras videotaped the license plate but couldn’t see inside the car; nor, he must have assumed, could anyone record him on the deserted road he finally reached where he was undoubtedly confident that he was safe. What he didn’t notice was the CALFIRE video camera placed on that very road to monitor for brush fires. It caught a car’s headlights matching his on its way to the site he had chosen to dump the body. There was no direct evidence of the murder he had committed, just circumstantial, tech-based evidence. A jury, however, convicted him in just a few hours.”

– Judith Coburn, “A Private Investigator on Living in a Surveillance Culture: Can we be forgotten anymore?Common Dreams. August 27, 2018.

Read Full Post »

“the big growth industry right now is totalitarianism.

that’s not some intense rhetorical statement or political slogan, it’s a fact by the definition of the word.

the rise of ‘smart’ technology–cities that monitor the motions of their citizens, cars that report to insurance companies what you’re talking about, thermometers that keep an eye on where people are in the house and run that through a third-party database, and most of all google (alphabet now actually, google is just a subsidiary) which is contending for the richest company in the world–are based off expanding a practice of total universal data control and “realtime behavior modification”–usually through negative incentivization: if you talk about driving dangerously, your car tells the insurance company who increases your rates. this is the purpose behind ‘gamifying’ structures as well, to build a video-game-like reward/punishment system into capitalism where one’s behavior is constantly supported or opposed. functionally, tech bros are imposing a bad videogame morality system onto real life. or, not “are imposing” so much as “have imposed”, given how such a thing already defines how we interact with each other online. it is formalized, incentivized totalitarianism, to a degree that the totalitarian governments of the past couldn’t imagine (this totalitarianism is decentralized, abstracted to trends and aesthetics and the logic of capital).

they are trying to push total complacency as we move into the era of climate collapse. and it’s happening so fast, so efficiently, and so differently than people expected, no one knows how to respond.”

Read Full Post »

“The necessary condition for the reign of the GAFA (Google, Apple, Facebook, Amazon) is that beings, places, fragments of the world remain without any real contact. Where the GAFA claim to be “linking up the entire world,” what they’re actually doing is working toward the real isolation of everybody. By immobilizing bodies. By keeping everyone cloistered in their signifying bubble.

The power play of cybernetic power is to give everyone the impression that they have access to the whole world when they are actually more and more separated, that they have more and more “friends” when they are more and more alone. The serial crowd of public transportation was always a lonely crowd, but people didn’t transport their personal bubble along with them, as they have done since smartphones appeared. A bubble that immunizes against any contact, in addition to constituting a perfect snitch.

This separation engineered by cybernetics pushes in a non-accidental way in the direction of making each fragment into a little paranoid entity, towards a drifting of the existential continents where the estrangement that already reigns between individuals in this “society” collectivizes ferociously into a thousand delirious little aggregates.

In the face of all that, the thing to do, it would seem, is to leave home, take to the road, go meet up with others, work towards forming connections, whether conflictual, prudent, or joyful, between the different parts of the world. Organizing ourselves has never been anything else than loving each other.”

comité invisible, “50 Nuances of Breakage,” from Now. 2017.

Read Full Post »

Older Posts »