Author: Jill Toh

Human-in-the-loop is not the magic bullet to fix AI harms

In many discussions and policy proposals related to addressing and fixing the harms of AI and algorithmic decision-making, much attention and hope has been placed on human oversight as a solution. This article by Ben Green and Amba Kak urges us to question the limits of human oversight, rather than seeing it as a magic bullet. For example, calling for ‘meaningful’ oversight sounds better in theory than practice. Humans can also be prone to automation bias, struggle with evaluating and making decisions based on the results of the algorithm, or exhibit racial biases in response to algorithms. Consequentially, these effects can have racist outcomes. This has already been proven in areas such as policing and housing.

Continue reading “Human-in-the-loop is not the magic bullet to fix AI harms” →

AI and its hidden costs

In a recent interview with The Guardian, Kate Crawford discusses her new book, Atlas AI, that delves into the broader landscape of how AI systems work by canvassing the structures of production and material realities. One example is ImageNet, a massive training dataset created by researchers from Stanford, that is used to test whether object recognition algorithms are efficient. It was made by scraping photos and images across the web and hiring crowd workers to label them according to an outdated lexical database created in the 1980s.

Continue reading “AI and its hidden costs” →

Racist and classist predictive policing exists in Europe too

The enduring idea that technology will be able to solve many of the existing problems in society continues to permeate across governments. For the EUObserver, Fieke Jansen and Sarah Chander illustrate some of the problematic and harmful uses of ‘predictive’ algorithmic systems by states and public authorities across the UK and Europe.

Continue reading “Racist and classist predictive policing exists in Europe too” →

Online hate and harassment continue to proliferate

A recent report by ADL, an anti-hate organisation in the US, has shown that social media platforms have consistently failed to prevent online hate and harassment. Despite the self-regulatory efforts made by social media companies, results from ADL’s annual survey shows that the level of online hate and harassment has barely shifted in the past three years. These online experiences disproportionately harm marginalised groups, with LGBTQI+, Asian-American, Jewish and African-American respondents reporting higher rates of various forms of harassment. Many of these problems are intrinsic to the ways in which the business models of social media platforms are optimised for maximum engagement, further exacerbating existing issues in society.

Continue reading “Online hate and harassment continue to proliferate” →

Online proctoring excludes and discriminates

The use of software to automatically detect cheating on online exams – online proctoring – has been the go-to solution for many schools and universities in response to the COVID-19 pandemic. In this article, Shea Swauger addresses some of the potential discriminatory, privacy and security harms that can impact groups of students across class, gender, race, and disability lines. Swauger provides a critique on how technologies encode “normal” bodies – cisgender, white, able-bodied, neurotypical, male – as the standard and how students who do not (or cannot) conform, are punished by it.

Continue reading “Online proctoring excludes and discriminates” →

Filtering out the “Asians”

The article’s title speaks for itself, “Your iPhone’s Adult Content Filter Blocks Anything ‘Asian’”. Victoria Song has tested the claims made by The Independent: if you enable the “Limit Adult Websites” function in your iPhone’s Screen Time setting, then you are blocked from seeing any Google search results for “Asian”. Related searches such as “Asian recipes,” or “Southeast Asian,” are also blocked by the adult content filter. There is no clarity or transparency to how search terms are considered adult content or not, and whether the process is automated or done manually. Regardless of intention, the outcome and the lack of action by Google or Apple is unsurprising but disconcerting. It is far from a mistake, but rather, a feature of their commercial practices and their disregard to the social harms of their business model.

Continue reading “Filtering out the “Asians”” →

Racist technology in action: Gun, or electronic device?

The answer to that question depends on your skin colour, apparently. An AlgorithmWatch reporter, Nicholas Kayser-Bril, conducted an experiment that went viral on Twitter, showing that Google Vision Cloud (a service which is based on a subset of AI known as “computer vision” that focuses on automated image labelling), labelled an image of a dark-skinned individual holding a thermometer with the word “gun”, whilst a lighter skinned individual was labelled holding an “electronic device”.

Continue reading “Racist technology in action: Gun, or electronic device?” →

Corporatespeak and racial injustice

In light of the Black Lives Matter protests in the U.S. and protests against police brutality in Europe, technology companies have been quick to release corporate statements, commitments, campaigns and initiatives to tackle discrimination and racial injustice. Amber Hamilton evaluated 63 public facing documents from major technology companies such as Facebook, Instagram, Twitter, YouTube, Airbnb and TikTok.

Continue reading “Corporatespeak and racial injustice” →

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

Up ↑