Advanced Search
Content Type: Examples
Human raters have played a significant role in the rapid improvement in the machine learning models that fuel modern AI. The raters evaluate the algorithmic output of search engines and AI chatbots and provide "Reinforcement Learning with Human Feedback" (RLHF) – the technical name for the deployment of such ratings to improve AI models. The efforts of these workers, who are mostly located in the global South but include thousands in the US, is downplayed by the technology companies to whom…
Content Type: Examples
In two cases brought by Worker Info Exchange and the App Drivers and Couriers Union on behalf of drivers, the Court of Appeal in Amsterdam has upheld a 2021 ruling in a lower court that under the GDPR Uber and Ola Cabs must disclose the personal information and profiling that the companies use to create "fraud probability scores" and "earnings profiles" to workers. These scores and profiles are in turn used in automated decision making to allocate work and pay rates. The court rejected the…
Content Type: Examples
A large-scale preprint study of more than 100 million rides between 2018 and 2019 in Chicago, where a 2020 law requires ride-hailing apps to disclose fares, finds that the dynamic pricing algorithms used by ride-hailing companies such as Lyft, Uber, and Via are socially biased. The finding is in line with earlier studies by other organisations such as the Princeton Review that found bias in algorithmic pricing. The researchers found that prices for rides varied according to the average…
Content Type: Examples
An eight-country study of Amazon employees has found that 57% say the company's performance monitoring system damages their mental health, 51% (65.7% of drivers) say it's had a negative effect on their physical health, and 59% feel the monitoring is excessive. In addition, 58% say Amazon doesn't explain clearly how it uses the data it collects on workers. Injury rates at Amazon warehouses are above the industry average.https://uniglobalunion.org/news/globalsurvey23/Publication: UNI Global…
Content Type: Examples
Workers in Amazon warehouses are tracked closely by a system that records every minute of "time off task" via the radio frequency handheld scanners workers use to track customer packages. Breaching strict time off task time limits can get an employee fired. Time off task includes bathroom breaks, talking to other Amazon employees, or going to the wrong floor of a warehouse: - managers may be required to ask offenders to account for each missing minute. https://www.vice.com/en/article/…
Content Type: Examples
AI-powered cameras made by the startup Netradyne and used in Amazon's delivery vans incorrectly penalises drivers for events beyond their control or which do not constitute unsafe driving such as if they are cut off by another vehicle. The data collected by the cameras is sent to Amazon, which uses the information to evaluate drivers' performance by assigning them with a score for safe driving. https://www.vice.com/en/article/88npjv/amazons-ai-cameras-are-punishing-drivers-for-mistakes-…
Content Type: Examples
A BBC Panorama investigation that brought hidden cameras into a UK-based Amazon warehouse found that workers walked up to 11 miles in a shift and had just 33 seconds on average to find each product, following instructions from a handset. Experts on stress at work say the warehouse conditions are ideal for increasing the risk of mental and physical illness.https://www.bbc.co.uk/news/business-25034598Publication: BBCPublication date: 2013-11-25Writer: BBC
Content Type: Examples
Humans who review footage of warehouse workers flagged by Amazon’s AI computer vision system to check for employee errors - are themselves surveilled in detail to ensure they make punishing targets. The workers, who are paid as little as £212 a month to review thousands of images and videos per day, report physical problems, deteriorating eyesight, and cognitive exhaustion. https://www.thebureauinvestigates.com/stories/2022-11-21/the-eyes-of-amazon-a-hidden-workforce-driving-a-vast-…
Content Type: Examples
The San Francisco-based company Emotiv claims its MN8 electroencephalography device can be worn comfortably for a full workday. The device monitors an employee's brainwaves with the goal of creating safer, more efficient workplaces by monitoring workers' brain signals for signs of stress or distraction. Company president and neuroscientist Oliver Oullier says the data the device collects makes it possible to reschedule tasks and working hours to lessen stress and increase focus. Emotiv says all…
Content Type: Examples
JP Morgan Chase's hundreds of thousands of employees are monitored in detail throughout their working day with the collected data sent to the data management system Workforce Activity Data Utility, which the company began building shortly before the coronavirus pandemic started. Some employees say it is not clear to them why the bank tracks how much time they spend on Zoom calls or writing emails or how this data will be used to assess their performance. They say that fears around how the data…
Content Type: Examples
Delivery drivers in Jakarta use GPS-spoofing apps in order to improve their chances of selection by the Gojek delivery and transport app, an equivalent to Apple Pay, Postmates, Venmo, and Uber all in one. Gojek that operates in more than 200 cities in Indonesia, Singapore, Vietnam, and Thailand. Other grey market apps enlarge details of orders that are too small to read, automate bidding, and apply filters to open orders. Some apps are distributed via Google Play; more are sold via driver…
Content Type: Examples
Content moderators working in the south Asian IT hub Hyderabad say their work reviewing still and video images of sexual and violent content is straining their mental health. Paid less than £8 a day, the moderators say the wellness coaches the company supplies do little beyond moving them to a less sensitive queue until they feel better. In some cases, managers tell the moderators to leave up content they'd rather remove, such as animal killings. Humans are needed for this work; AI is…
Content Type: Examples
Police in Western Australia have demanded that ABC hand over all the footage of climate protesters it collected in preparing a programme ("Four Corners") about them. More than 40 civil society groups have opposed the request, saying that the demand undermines press freedom and urging the broadcaster to protect its journalists' sources. https://www.theguardian.com/australia-news/2023/oct/05/wa-police-order-abc-to-hand-over-four-corners-footage-of-woodside-protest?cid=…
Content Type: Examples
Myanmar's ruling junta has begun a pilot census in 20 townships across the country, and has said that a national census needs to be completed in 2024 before new elections take place, which could be in 2025. Critics warn that the census will be used to increase surveillance of opponents, including protesting civil servants, doctors, and teachers. Myanmar also intends to roll out an electronic identification system and has asked China for help in designing it.https://www.abc.net.au/news/2023-10-…
Content Type: Examples
The Israeli government is bringing forward a bill that would permit the police to place facial recognition cameras in public places, including at events such as protests, as long as a police officer is convinced that the cameras' operation does not present undue invasion of any individual's privacy and the purpose is to prevent, thwart, or detect serious crimes, locate missing persons, and enforce restraining orders or bans. The bill is based on another pending bill that is meant to cover a…
Content Type: Examples
The Taliban could repurpose a plan devised by the Americans before their 2021 departure for a four-year programme to create a large-scale camera network to surveil Afghan cities, The capital, Kabul, already has thousands of cameras. The Taliban administration has also consulted with Huawei about potential cooperation. Rights groups are concerned the technology will be used to crack down on protesters and some analysts question the regime's ability to pay for the programme, but records of…
Content Type: Examples
Based on a facial recognition match, the New York City Police Department sent more than 50 officers to besiege the home of a racial justice organiser, claiming he had shouted in an officer's ear at a protest in the summer of 2020. The officers were unable to produce a warrant when asked, but deployed snipers, drones, helicopters, and police dogs. The organiser, Derrick Ingram, has filed a lawsuit.https://www.the-independent.com/news/world/americas/crime/facial-recognition-technology-police-…
Content Type: Examples
The three Democratic members of the five-member Privacy and Civil Liberties Board, an independent agency within the federal branch of the US government, have recommended that the FBI and other government agencies should be required to obtain the approval of a court before reviewing the communications of US citizens collected under Section 702 of the Foreign Intelligence Surveillance Act. The Board nonetheless recommended that the programme be renewed when it comes up for review at the end of…
Content Type: Examples
A BBC investigation found more than 150 cases in which UK police abused their body-worn cameras by deleting footage, turning them off when using force, and sharing videos on WhatsApp, other social media, or in person. The cameras, which have cost at least £90 million over the last decade, were intended to improve transparency, protect police against malicious complaints, and benefit victims by improving the quality of evidence. In some cases, the behaviour was said to be unlawful by the…
Content Type: Examples
During the runup to the 2023 general election, the Zimbabwean government sent drones to monitor a rally organised by the Citizens Coalition for Change. The country also uses drones to monitor motorists, control the border, and limit crime; officials in the Kavango Zambezi Transnational Park have used it to complete an elephant census. Law enforcement authorities are now adopting facial recognition to help combat more sophisticated crimes. Zimbabwe's opposition fears that the technologies could…
Content Type: Examples
In the year after protests began in Iran in September 2022, Telegram has emerged as the social medium of choice for both the protesters and the regime they oppose. The Iranian authorities have been able to use Telegram to identify and shame protesters and broadcast false confessions, as well as mount disinformation campaigns. In September 2023, the authorities increased penalties for violations of the hijab law.https://www.codastory.com/newsletters/iran-protests-anniversary-censorship-…
Content Type: Examples
The rise of hybrid work has led to a rise in "bossware": increasingly intrusive technology that monitors employees, tracks their locations, and watches or listens to office workers via cameras and microphones. 90% of such systems can give employers a list of everything a worker has done that day. The cost of such systems has dropped, as has employer trust in staff. The increasing surveillance, now with AI predictive functions, threatens job security and increases the power companies have over…
Content Type: Examples
Drivers for app-based companies like Uber, tired of their lack of transparency, share their experience and swap tips to help each other game the platforms to their advantage via in-person workshops and Telegram groups, aided by the Indian Federation of App-based Transport Workers and the Telangana Gig and Platform Workers' Union. Similar movements exist around the world. In 2021, a Dutch court upheld a complaint by Uber and Ola drivers from the UK and Portugal asking those companies to provide…
Content Type: Examples
Behind every powerful AI system are huge numbers of people labelling and clarifying data to train it, contracted by companies like Remotasks, a subsidiary of Silicon Valley-based data vendor Scale AI, whose customers include the US military and OpenAI. Often the workers, who are assigned tasks they don't understand for a purpose they don't know, are sworn to secrecy. Yet labelling is crucial; it can make the difference between a car stopping to spare the person walking a bike across the road or…
Content Type: Examples
More than 150 workers employed by third-party outsourcing companies to provide content moderation for AI tools on Facebook, TikTok, and ChatGPT depend have pledged to create the African Content Moderators Union. The move to create such a union began in 2019 when the outsourcing company Sama fired Facebook content moderator Daniel Motaung for trying to form a union. https://time.com/6275995/chatgpt-facebook-african-workers-union/Publication: TimePublication date: 2023-05-01Writer: Billy…
Content Type: Examples
Outsourced artists, designers, copywriters, software developers, and call centre operators in the global south are the first to feel the effects of the arrival of generative AI, as client companies see the new technology as a way of cutting costs. Some workers are adding prompt engineering to their advertised skillsets or offer a service copy-editing and fact-checking AI-generated output; others say that using AI tools helps them produce more work faster, albeit for less money. Outsourcing…
Content Type: Examples
Technology companies that call themselves "AI first" rely on heavily surveilled gig workers who label data, deliver packages, moderate content, and perform gig work via platforms. Startups pressured by their venture capital funders even hire humans to pretend to be chatbots so they can claim to be "AI" companies. For these reasons, worker exploitation needs to be a central part of the discussion of the ethical development and deployment of AI systems.https://www.noemamag.com/the-exploited-labor…
Content Type: Examples
In a legal action, the Independent Workers Union of Great Britain and the App Drivers and Couriers Union claim that Uber's use of facial recognition software for its Real-Time ID Check to verify the identity of drivers is discriminatory because facial recognition software is known to be less accurate at identifying people with darker skin. The action was brought on behalf of two drivers whose accounts were terminated following errors made by the Microsoft-supposed facial recognition software.…
Content Type: Examples
A former Amazon warehouse worker writes that every day was "brutal" because of the "exploitative and dangerous" standards enforced by Amazon executives. Amazon's anxiety-inducing policies about bathroom use and low pay should be seen in context with fast food and retail workers, who frequently encounter violence on the job and many essential workers' struggle to afford the basic necessities of life. In response, workers are beginning to target investors as an important voice that can help…
Content Type: Examples
An excerpt from the new book "Arriving Today: From Factory to Front Door—Why Everything Has Changed About How and What We Buy", describes in detail the tracking systems used in Amazon warehouses to ensure workers meet their managers' targets. The system is a mix of surveillance, measurement, psychology, targets, incentives, slogans, and proprietary technologies.https://www.wsj.com/articles/the-way-amazon-uses-tech-to-squeeze-performance-out-of-workers-deserves-its-own-name-bezosism-…