Women currently make up just 20% of employees in technical roles at major machine learning companies, and that only 12% of AI researchers and 6% of professional software developers are women. This paper presents the case for feminist technology diplomacy as a pathway to integrating gender and feminist perspectives into foreign policy approaches to AI.
 

By: Elise Stephenson and Isobel Barry

Posted on 17 September 2024

Key findings

 

Artificial Intelligence (AI) has rapidly gained global momentum, with technological developments spurred on by generative AI and increased automation across military, medical, humanitarian, and educational domains, hastening the need to prioritise good international governance. To realise the gains whilst mitigating the risks of AI, robust governance frameworks are required to which feminist foreign policy (FFP) approaches are relevant. 

AI regulation that focuses on mitigating harms toward minoritised groups, particularly women, should be a critical priority for governments and international organisations. Failing to address these issues risks disenfranchising half the global population, ultimately leading to missed opportunities for gains in productivity and efficiency driven by AI advancements. 

This paper presents the case for feminist technology diplomacy as a pathway to integrating gender and feminist perspectives into foreign policy approaches to AI. Through this research, we developed some key feminist AI principles, including...

  1. Acknowledging AI is gendered (and raced, etc.): Whilst AI technologies are often depicted as ‘objective’ and technologists as ‘neutral’, emerging technologies are fundamentally cultural and shaped by their context.
  2. Recognising that technology alone is not the solution: We also need social scientific and creative critical thinking. 
  3. Reconciling that some feminists are at odds with some uses of AI: e.g. military AI technologies, predictive policing. 
  4. Distinguishing bias as (often) intentional, correction must be proactive: Bias is not always or even often an accidental by-product or technical error, but rather, a reinforcement of existing power relationships. 
  5. Moving beyond critique to practice: Merely identifying issues is not enough, we also need to focus on developing inherently feminist AI technologies.

Contact