AI Justice: An Intersectional Feminist Approach to Artificial Intelligence

Value of the Information & Strength of the Argument

The value of the information lies in its critical perspective on AI, drawing on feminist and decolonial theories to highlight power dynamics often overlooked in mainstream AI discourse. The argumentation is coherent and well-structured, systematically building from the non-neutrality of technology to the colonial nature of Big Tech and the local impacts of global technologies. Nony supports her claims with concrete examples, such as the sexualization of generated images and the case of Facebook’s free data initiative, which illustrate the real-world consequences of biased AI. However, the argumentation is primarily theoretical and relies heavily on the authority of the speaker and cited scholars, without providing detailed empirical evidence or systematic citations during the talk. The call for community-rooted and emancipatory AI is compelling but remains at a high level, lacking concrete implementation strategies.

Scientific Rigor, Source Quality, Title Accuracy

The talk demonstrates a strong engagement with feminist and decolonial scholarship, referencing key thinkers and concepts such as Joy Buolamwini’s ‘coded gaze’ and technocolonialism. However, the presentation does not systematically cite sources, and the speaker mentions that references will be sent in a follow-up email, which limits the immediate verifiability. The title accurately reflects the content, which is an expert opinion rather than a systematic review or original study. The seminar format allows for interactive discussion, but the lack of formal citations and the informal tone reduce the scientific rigor. The speaker’s expertise is evident, but the reliance on personal authority and the absence of a clear methodology for evaluating claims weaken the overall reliability.

The title accurately reflects the content, which focuses on an intersectional feminist and anti-colonial approach to AI.

7/10

The speaker is a senior researcher with expertise in philosophy of technology and political theory of AI. The talk is an expert opinion grounded in feminist and decolonial scholarship, but it lacks systematic citation of sources during the presentation. The arguments are coherent and well-structured, but the lack of verifiable references and the informal seminar format reduce the overall reliability.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 00:00Introduction and welcome, setting the seminar's purpose and inviting participants to share their location and skills.
• 10:00Introduction of the first pillar: technologies and sciences are not neutral, and the need to reinvestigate socially constructed systems.
• 20:00Discussion of the 'coded gaze' concept from Joy Buolamwini's work, illustrating bias in AI vision technologies.
• 30:00Introduction of the second pillar: Big Tech as colonial politics by other means, including algorithmic colonization and technocolonialism.
• 40:00Examples of technocolonialism, such as Facebook's free data initiative, and the impact on dependency and access.
• 50:00Discussion of the third pillar: global technologies affect life matters locally, with examples of media ownership and biometric systems.
• 1:00:00Call for community-rooted and emancipatory AI, and the need for collective action and insurrection against oppressive technologies.
• 1:10:00Closing remarks and opening the floor for questions and discussion.

• Anaïs Nony's personal website — Speaker's profile and research focus.
• JIAS seminar registration form — Registration for the next seminar in the series.
• Johannesburg Institute for Advanced Study (JIAS) website — Information about the institute and its activities.

• Joy Buolamwini's research on facial recognition bias — The 'coded gaze' concept and algorithmic bias in facial recognition.
• Article on technocolonialism — Academic discussion of technocolonialism and digital dependency.

• Critique of AI ethics frameworks — Some scholars argue that mainstream AI ethics frameworks, including feminist approaches, may be co-opted by industry and fail to address structural power imbalances.