AI in Mexico(?) ...towards community-led practices

Mexico had an AI policy strategy drafted back in 2018. With it, the Mexican government prided itself in being one of the 10 countries in the world to have such a strategy which is heavily focused on the readiness of the country, responsible innovation with some hints of civil society involvement in the future and a page on ethics.
My worry is that this goes with the dominant wave of “AI ethics washing” focused on bias and fairness, neglecting any questions regarding systemic injustice, power structures and control of infrastructures in the country. I believe that when elemental systemic issues remain untouched, no amount of great innovations, bias and fairness training would be able to encapsulate and solve the complexity of the racial, power and control structures in Mexico.

So, moving from the mere knowledge that ‘technical’ systems actually involve ‘social’ systems, we can then be on a path of understanding that the social context in which these ‘social’ ‘technical’ (“Sociotechnical”) systems operate are actually interdependent.

Mexico has and continues to struggle with its racism, a country that has tried its best at denying and oppressing indigenous peoples, ignoring and marginalizing the mexican Black communities and peasant and rural communities. Where racism is masked and denied as classism and remains a taboo. Then you wonder, “how can we even consider our AI systems to be fair and just?” and I think good intentions dressed as ‘AI for good’ ‘ethical and bias trainings’ could simply not be enough.

Of course, there is the role of the geopolitical power imbalances that colonialism purposely designs and actively shapes AI governance and the debates going around. This obviously, has played out in the recent rush to form global AI ethics guidelines where developing countries in Africa, Latin America, and Central Asia have been largely left out of the discussions resulting in: developed countries continuing to disproportionately benefit from global norms shaped for their advantage while at the same time colonizing structures within developing countries in LatAm (like Mexico) remain, then marginalized communities will continue to be oppressed.

In this vein, then how can Mexico and Latin America achieve decolonial and ‘just AI’, if at all possible?

I am not against Algorithmic based models, I think they are a good way to expand the imagination and understand a problem further, even discover new things in the natural sciences, however when used in social contexts, there is a whole complexity around them that no algorithm can nor should fix on its own or at all.

Right now, there are already some ongoing projects, ideas, visions in Mexico reimagining a “world where all worlds fit”. Indigenous communities and their involvement in digital activism are using the internet to bring their narratives to the digital world. Projects such as translating platforms for some indigenous languages. The UNAM (National Autonomous University of Mexico) sessions on ‘rewriting technologies’ are doing some extraordinary work on seeing technology with new visions towards community-led practices, with situated and a critical approach to technology development in Mexico, a much needed initiative I think, to listen to all voices in Mexico (especially those that are constantly left out) hear their perspectives and build with them not despite them.

Back in 2018 while preparing my final Master’s project (which was quite small scale) and seeing the various ways my own community and other rural communities around have been using their own resources to build their own resilience, social capital and overall empowerment in response to the failed policies on high securitization and militarization of the country towards “fighting” crime, a project with collective action was born... As these communities were the driving force of this project, centering these community-led practices where empowerment and collective effort are found was vital.

In a country where the narrative has been driven by those in power, the stigma against marginalized communities continues to be a driving force for further marginalization and securitization of their spaces. Many of these communities are already making use of digital technologies to write their own narratives, to raise the alarm on the injustices, to change the narrative online.

So the purpose of this project was to point out the already quite obvious but not taken serious alternative to the ongoing violence in the country — comunity-led practices. In this way, and working with communities present online, the idea was to bring their data together of all the actionable, alternative systems they have in place with the goal of dismantling the state’s oppressive, militarized failed policy and to make the point that the state’s policies and funding could and should be diverted towards the demands of these communities.

These communities already have processes in place that work for them and empower their communities. And it makes you think that perhaps the state’s approach should be not of what they can do without consulting these communities, but what they should do to support them and empower them with their knowledge and systems already in place or work towards scrapping their failed policies and design new policies taking into account the contexts of these communities.

Perhaps my proximity to the context of many of these communities like my community allowed me to be context aware of any technical development and research involved. But I really believe that as researchers/developers our job is not that of an expert but an accomplice/a facilitator in supporting, facilitating and organizing where the communities have tutelage over these processes. But of course you wonder, would these methods take us to liberation? I would like to clarify that while this may look like studying a social movement, that is not the goal. The goal is to collect as much data as possible on all the ways the government (police, military, etc) have failed these communities time and time again. “Policing the government... “

While I was really excited to see Cathy O’Neills article on ‘algorithms for defunding the police' a few months ago (which this project somehow intends to do in the mexican context) and as I go into a PhD and further develop this project with my community and other rural communities in Mexico. Constant reflection and active participation with the communities is an essential part of how I see my involvement. However, as a member of one of those communities but also a recurrent inhabitant, I ask myself “should I even be involved at all?” “How I as an insider and an outsider should (“if”) I should be involved and why?”

So my hopes for the present and future of AI and technology in general is as Sasha Costanza-Chock mentions in her book Design Justice, to “continually challenge the ways that inequality is reproduced through the design of algorithms and support systems and emphasize the development of communities as an active participative process”. ## To reimagine new paradigms and rewrite technology.


Costanza-Chock, S. (2020) Design Justice: Community-led practices to build the worlds we need. The MIT Press.

Ellis, B.H. and Abdi, S.M. (2017) Building community resilience to violent extremism through genuine partnerships. The American psychologist, 72(3), 289-300.

Escobar, A. (2011). Sustainability: design for the pluriverse. Development, 54(2), 137–140. Hao, K. (2019) in 2020, let’s hope AI ethics-washing and actually do something. Technology Review. Ai ethical choices. Available at: Accessed on 20/07/2020.

Kind, C. (2020) The term ‘ethical AI’ is finally starting to mean something. Venture Beat. The Machine: making sense of AI. Available at: Accessed on 24/08/2020.

Strategy AI-MX – Martinho-Trustwell, E., Miller, H., Asare, I. N, Petheram, A., Stirling R., Mont, G.C., Martinez, C. (2018) Towards and AI strategy in Mexico: Harnessing the AI revolution. White Paper.

Mohamed, S., Png, M. & Isaac, W. (2020) Decolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligence. Philosophy and Technology

— Yadira