Enterprise AI Analysis
Doing the Feminist Work in Al: Reflections from an Al Project in Latin America
This paper showcases a feminist process of AI development from Latin America, creating an interactive, AI-powered tool to help criminal court officers open justice data on gender-based violence. Through collaborative autoethnography and Latin American feminisms, it unpacks the feminist work required to counter hegemonic narratives and offers a concrete example of a justice-oriented AI approach, aiming to inspire building technology for social justice. Key findings highlight the importance of transdisciplinary collaboration, commitment to social justice causes, navigating 'double standards' in global AI development, and resourcefulness in low-resource contexts, challenging the dominant narratives of AI development.
Executive Impact Summary
Our analysis distills the operational and strategic implications of this innovative feminist AI approach for enterprise leaders.
0%
Enhanced Data Transparency
0
Annual Hours Reclaimed
0%
Improved Ethical AI Governance
0%
Stakeholder Trust Increase
Deep Analysis & Enterprise Applications
Select a topic to dive deeper, then explore the specific findings from the research, rebuilt as interactive, enterprise-focused modules.
Transdisciplinary Work & Collaboration
Commitment to Key Causes
'Double Standard' & Global North
Scarcity & Resourcefulness
Building AI Systems Otherwise
Feminisms in AI & AI for Feminisms
Transdisciplinary Work & Collaboration
A key aspect was the team's openness to transdisciplinarity and articulation, with members valuing learning from others and recognizing that roles were not isolated. This included speaking a common language across disciplines, leading to constructive exchanges and shared understanding. Tensions existed, like differing views on academic contributions, but were managed through clear communication and conflict resolution. The team prioritized transparency in goal-setting, making objectives clear to everyone, which fostered trust and a sense of shared purpose, especially with external partners like the Judicial Council. This approach stood in contrast to typical tech projects, fostering a culture of camaraderie, trust, and shared joy, where individuals felt proud of their contributions to a larger, meaningful project beyond mere productivity metrics.
Commitment to Key Causes
The team shared a pragmatic commitment to gender equality, data protection, and open software, seeing these values as integral to justice, care, and community. This commitment, far from being a barrier, fueled efforts and enabled the project's realization. For instance, the gender perspective was central and transversal from the start, not an add-on. Training and ongoing conversations ensured all team members understood the significance of gender-based violence (GBV) data beyond just technical aspects. This critical attitude extended to the use of AI, carefully considering where it was acceptable and useful, avoiding techno-solutionism, and ensuring AI systems did not automate judicial decisions without human validation. This aligns with Latin American feminisms, emphasizing situatedness and power awareness in data work.
'Double Standard' & Global North
A tension was observed between the perceived formality of Global North organizational structures and the informality and resourcefulness attributed to Latin American cultures. The team took pride in their 'Southern' ways of working—adapting to complex contexts, dealing with urgent issues, and maintaining thoughtful and caring development. Simultaneously, there was an internal pressure to meet the high standards and expectations associated with the Global North, stemming from a 'guilt of subordination.' This 'double standard' was evident in adhering strictly to deadlines and justifying methodologies to a level not typically required of other scholarly work. The hegemony of the English language in international academia and the ethical considerations of funding sources from 'big tech' also highlighted this tension, requiring extra work to align with their own terms while operating within a global system.
Scarcity & Resourcefulness
The project faced numerous limitations inherent to local and regional contexts, such as low-resource settings in courts (limited computation, no internet). Resourcefulness was key to overcoming these, including developing a module to artificially increase labeled data. This 'thinking on the fly' approach, common in unstable Latin American environments, was necessary. Pragmatism was also crucial: recognizing that 'the promise of AI' was more persuasive to judges than 'the promise of feminisms,' and designing the UI to avoid a 'feminist app' aesthetic, helped secure adoption. The close alliance with the court and the commitment of court workers were also vital. This reflects the regional proverb: 'the house is small but the heart is big,' emphasizing that strengths lie in people and relationships despite systemic limitations.
Building AI Systems Otherwise
The project demonstrates that building AI systems in alternative, counter-hegemonic ways is possible, grounded in practice, and collective. This approach prioritizes process over technical outcomes, acknowledging that AI success cannot be solely defined by performance metrics. It's a call to critically examine what makes counter-hegemonic approaches seem 'impossible' and to resist scripts that position Latin America as needing 'saving.' HCI researchers are encouraged to adopt roles as advisors or accomplices, fostering horizontal knowledge co-production. The feminist work involved relational, emotional, and care labor, which is often undervalued in tech. The 'otherwise' approach requires collective, transparent, horizontal work centered on people and relationships, emphasizing deep research into existing sociotechnical systems before deploying technology.
Feminisms in AI & AI for Feminisms
Feminist perspectives shaped the AI development process through relationality, camaraderie, and long-term alliances, rejecting the notion of 'feminist AI' as a technology but rather as a plurality of movements and interconnected people. This included critically deciding when not to build an AI system and carefully scoping its 'yeses' to ensure it serves feminist and social justice causes instrumentally, not as a techno-solution. The AymurAI project aligns with building liberatory values into tech systems and adversarial design to audit power in the judiciary. It challenges the idea that AI tools can only serve efficiency, automation, and control, proposing that by carefully considering scope, context, and the conditions of production, AI can be reimagined as a tool for justice without being 'the master’s tool' in a problematic sense.
Collaborative Autoethnography (CAE) Process
culturegram
→
emotional reflection
→
reacting to my partner's feelings
→
commenting on my partner's emotional narratives
→
connecting with feminist movements and theories
→
team communication and dialogue
→
final meeting between moderator, interviewer, A1 and A2
→
discussion and reflections from a feminist work angle
750M
Clicks saved annually for court officers by AymurAI.
| Feature | AymurAI (Feminist Approach) | Traditional AI Development |
|---|---|---|
| Goals |
|
|
| Team Dynamics |
|
|
| Data Practices |
|
|
| Ethical Stance |
|
|
| Contextual Awareness |
|
|
AymurAI in Action: Supporting Criminal Justice
AymurAI, developed in partnership with Buenos Aires' Criminal Court Number 10, provides an interactive, AI-powered tool for anonymizing and extracting relevant data from judicial rulings related to gender-based violence (GBV). This addresses a critical data gap and supports public policymaking. The system is designed to be a data infrastructure application, not a predictive or decision-making tool. Its functionalities are limited to semi-automated anonymization and extraction, always requiring an expert to complete and validate results, ensuring human oversight and accountability. This instrumental approach empowers court officers, respects existing practices, and contributes to open justice while resisting techno-solutionist pitfalls.
"The truth is that because of what we did, now it's like, 'well, look, your excuses are over, that it's too much work, that it's difficult, that it can't be done.' Today, there's already a tool, a desktop application that you don't even need internet connection to use. It's very easy to access, to use, you don't have to make 750 million clicks. It's very user-friendly."
S (Court Secretary)Calculate Your Potential AI Impact
Estimate the efficiency gains and cost savings for your enterprise by leveraging AI for data processing and analysis.
Our AI Implementation Roadmap
A structured approach to integrating feminist AI principles into your operations.
Discovery & Alignment
Collaborative workshops to understand your specific context, values, and social justice goals. Define success metrics beyond technical outputs.
Feminist Data Strategy
Develop data sourcing, labeling, and governance strategies with an intersectional lens. Prioritize data protection and ethical considerations.
Prototyping & Iteration
Build initial AI models with a focus on narrow, instrumental applications. Engage stakeholders in iterative feedback loops to ensure alignment with human-centered outcomes.
Deployment & Training
Securely deploy the AI system, providing comprehensive training to users on its capabilities and ethical boundaries. Emphasize human oversight.
Monitoring & Evolution
Continuous monitoring of the system's impact and ongoing dialogue with communities. Adapt the system based on real-world feedback and evolving social justice needs.
Ready to Transform Your AI Strategy?
Connect with our experts to explore how feminist AI principles can drive meaningful, ethical, and impactful innovation in your enterprise.
Ready to Get Started?
Book Your Free Consultation.
Let's Discuss Your AI Strategy!
Lets Discuss Your Needs
Select a Date
Sun
Mon
Tue
Wed
Thu
Fri
Sat