Machine Learning and the Reproduction of Inequality

Machine learning is the process behind increasingly pervasive and often proprietary tools like ChatGPT, facial recognition, and predictive policing programs. But these artificial intelligence programs are only as good as their training data. When the data smuggle in a host of racial, gender, and other inequalities, biased outputs become the norm.

By Catherine Yeh and Sharla Alegria for SAGE Journals on November 15, 2023

This is how AI image generators see the world

Artificial intelligence image tools have a tendency to spin up disturbing clichés: Asian women are hypersexual. Africans are primitive. Europeans are worldly. Leaders are men. Prisoners are Black.

By Kevin Schaul, Nitasha Tiku and Szu Yu Chen for Washington Post on November 20, 2023

AI is nog lang geen wondermiddel – zeker niet in het ziekenhuis

Tumoren ontdekken, nieuwe medicijnen ontwikkelen – beloftes genoeg over wat kunstmatige intelligentie kan betekenen voor de medische wereld. Maar voordat je zulk belangrijk werk kunt overlaten aan technologie, moet je precies snappen hoe die werkt. En zover zijn we nog lang niet.

By Maurits Martijn for De Correspondent on November 6, 2023

AI is bevooroordeeld. Wiens schuld is dat?

Ik ga in gesprek met Cynthia Liem. Zij is onderzoeker op het gebied van betrouwbare en verantwoorde kunstmatige intelligentie aan de TU Delft. Cynthia is bekend van haar analyse van de fraudedetectie-algoritmen die de Belastingdienst gebruikte in het toeslagenschandaal.

By Cynthia Liem and Ilyaz Nasrullah for BNR Nieuwsradio on October 20, 2023

Use of machine translation tools exposes already vulnerable asylum seekers to even more risks

The use of and reliance on machine translation tools in asylum seeking procedures has become increasingly common amongst government contractors and organisations working with refugees and migrants. This Guardian article highlights many of the issues documented by Respond Crisis Translation, a network of people who provide urgent interpretation services for migrants and refugees. The problems with machine translation tools occur throughout the asylum process, from border stations to detention centers to immigration courts.

Continue reading “Use of machine translation tools exposes already vulnerable asylum seekers to even more risks”

Al Jazeera asks: Can AI eliminate human bias or does it perpetuate it?

In its online series of digital dilemmas, Al Jazeera takes a look at AI in relation to social inequities. Loyal readers of this newsletter will recognise many of the examples they touch on, like how Stable Diffusion exacerbates and amplifies racial and gender disparities or the Dutch childcare benefits scandal.

Continue reading “Al Jazeera asks: Can AI eliminate human bias or does it perpetuate it?”

­Data Work and its Layers of (In)visibility

No technology has seemingly steam-rolled through every industry and over every community the way artificial intelligence (AI) has in the past decade. Many speak of the inevitable crisis that AI will bring. Others sing its praises as a new Messiah that will save us from the ails of society. What the public and mainstream media hardly ever discuss is that AI is a technology that takes its cues from humans. Any present or future harms caused by AI are a direct result of deliberate human decisions, with companies prioritizing record profits, in an attempt to concentrate power by convincing the world that technology is the only solution to societal problems.

By Adrienne Williams and Milagros Miceli for Just Tech on September 6, 2023

Vooral vrouwen van kleur klagen de vooroordelen van AI aan

Wat je in zelflerende AI-systemen stopt, krijg je terug. Technologie, veelal ontwikkeld door witte mannen, versterkt en verbergt daardoor de vooroordelen. Met name vrouwen (van kleur) luiden de alarmbel.

By Marieke Rotman, Nani Jansen Reventlow, Oumaima Hajri and Tanya O’Carroll for De Groene Amsterdammer on July 12, 2023

Racist Technology in Action: Stable Diffusion exacerbates and amplifies racial and gender disparities

Bloomberg’s researchers used Stable Diffusion to gauge the magnitude of biases in generative AI. Through an analysis of more than 5,000 images created by Stable Diffusion, they have found that it takes racial and gender disparities to extremes. The results are worse than those found in the real world.

Continue reading “Racist Technology in Action: Stable Diffusion exacerbates and amplifies racial and gender disparities”

Attempts to eliminate bias through diversifying datasets? A distraction from the root of the problem

In this eloquent and haunting piece by Hito Steyerl, she weaves the ongoing narratives of the eugenicist history of statistics with its integration into machine learning. She elaborates why the attempts to eliminate bias in facial recognition technology through diversifying datasets obscures the root of the problem: machine learning and automation are fundamentally reliant on extracting and exploiting human labour.

Continue reading “Attempts to eliminate bias through diversifying datasets? A distraction from the root of the problem”

Racist Technology in Action: Image recognition is still not capable of differentiating gorillas from Black people

If this title feels like a deja-vu it is because you most likely have, in fact, seen this before (perhaps even in our newsletter). It was back in 2015 that the controversy first arose when Google released image recognition software that kept mislabelling Black people as gorillas (read here and here).

Continue reading “Racist Technology in Action: Image recognition is still not capable of differentiating gorillas from Black people”

On Race, AI, and Representation Or, Why Democracy Now Needs To Redo Its June 1 Segment

On June 1, Democracy Now featured a roundtable discussion hosted by Amy Goodman and Nermeen Shaikh, with three experts on Artificial Intelligence (AI), about their views on AI in the world. They included Yoshua Bengio, a computer scientist at the Université de Montréal, long considered a “godfather of AI,” Tawana Petty, an organiser and Director of Policy at the Algorithmic Justice League (AJL), and Max Tegmark, a physicist at the Massachusetts Institute of Technology. Recently, the Future of Life Institute, of which Tegmark is president, issued an open letter “on all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4.” Bengio is a signatory on the letter (as is Elon Musk). The AJL has been around since 2016, and has (along with other organisations) been calling for a public interrogation of racialised surveillance technology, the use of police robots, and other ways in which AI can be directly responsible for bodily harm and even death.

By Yasmin Nair for Yasmin Nair on June 3, 2023

GPT detectors are biased against non-native English writers

The rapid adoption of generative language models has brought about substantial advancements in digital communication, while simultaneously raising concerns regarding the potential misuse of AI-generated content. Although numerous detection methods have been proposed to differentiate between AI and human-generated content, the fairness and robustness of these detectors remain underexplored. In this study, we evaluate the performance of several widely-used GPT detectors using writing samples from native and non-native English writers. Our findings reveal that these detectors consistently misclassify non-native English writing samples as AI-generated, whereas native writing samples are accurately identified. Furthermore, we demonstrate that simple prompting strategies can not only mitigate this bias but also effectively bypass GPT detectors, suggesting that GPT detectors may unintentionally penalize writers with constrained linguistic expressions. Our results call for a broader conversation about the ethical implications of deploying ChatGPT content detectors and caution against their use in evaluative or educational settings, particularly when they may inadvertently penalize or exclude non-native English speakers from the global discourse.

By Eric Wu, James Zou, Mert Yuksekgonul, Weixin Liang and Yining Mao for arXiv.org on April 18, 2023

Proudly powered by WordPress | Theme: Baskerville 2 by Anders Noren.

Up ↑