Decolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligence

Mohamed, Shakir; Png, Marie-Thérèse; Isaac, William

doi:10.1007/s13347-020-00405-8

Cited by 271 publications

(210 citation statements)

References 103 publications

Supporting

Mentioning

193

Contrasting

Unclassified

Order By: Relevance

“…Individuals who use this method would still need to communicate the validity of their conclusions to other people who hold a different set of reasonable credences. 17 There is some concern that this phenomenon is already occurring within the field of AI (Ricaurte 2019;Mohamed et al 2020).…”

Section: In the Absence Of Moral Agreement Is There A Fair Way To Dementioning

confidence: 99%

“…Although this is not their primary purpose, they could, in principle, serve as a locus for AI alignment.. 14 This ethic would have much in common with the notion of Ubuntu found in philosophical traditions in sub-Saharan Africa(Metz 2007). Ubuntu philosophy builds upon the idea that 'a person is a person through other persons' and that respect for these relationships, both inter-personally and as mediated through the technologies we use, is a key component of human dignity(Mhlambi 2020;Mohamed et al 2020). Important connections can also be draw between this principle and a feminist ethic of care (Gilligan 1993; Van Wynsberghe 2013).…”

mentioning

confidence: 99%

See 1 more Smart Citation

Artificial Intelligence, Values, and Alignment

2020

View full text Add to dashboard Cite

This paper looks at philosophical questions that arise in the context of AI alignment. It defends three propositions. First, normative and technical aspects of the AI alignment problem are interrelated, creating space for productive engagement between people working in both domains. Second, it is important to be clear about the goal of alignment. There are significant differences between AI that aligns with instructions, intentions, revealed preferences, ideal preferences, interests and values. A principle-based approach to AI alignment, which combines these elements in a systematic way, has considerable advantages in this context. Third, the central challenge for theorists is not to identify ‘true’ moral principles for AI; rather, it is to identify fair principles for alignment that receive reflective endorsement despite widespread variation in people’s moral beliefs. The final part of the paper explores three ways in which fair principles for AI alignment could potentially be identified.

show abstract

Section: In the Absence Of Moral Agreement Is There A Fair Way To Dementioning

confidence: 99%

mentioning

confidence: 99%

Artificial Intelligence, Values, and Alignment

2020

View full text Add to dashboard Cite

show abstract

“…Through basic textual analysis of Building an AI World 11 , Rebooting Regulation 12 , and Ethically Aligned Design 13 , there appears to be limited inclusion of thought leaders in groups beyond dominant technical, legal, psychological and economic backgrounds. Little reference is given to contributions by historically underrepresented communities, researchers and practitioners who have already evoked many of the themes explored in this paper involving sociotechnical coproduction, situatedness and critical reflexivity (Gasparotto 2016;Winchester III 2019;Mohamed et al 2020). It is here that the significant underrepresentation of Black, Indigenous, and racialized persons, women in particular, is once again apparent.…”

Section: Fitting In Regulationmentioning

confidence: 93%

Coproduction, Ethics and Artificial Intelligence

Govia

2020

jdsr

View full text Add to dashboard Cite

Over the past five years, artificial intelligence (AI) has been endorsed as the technical underpinning of innovation. Sensationalist representations of AI have also been accompanied by assumptions of technological determinism that distract from the ordinary, sometimes unassuming consequences of interaction with its systems and processes. Drawing on scholarship from cultural anthropology, along with science and technology studies (STS), this paper examines coproduction in a Canadian AI research and development context. Through interview responses and field observations it presents sites of sociotechnical entanglement and ethical discussion to highlight potential spaces of mediation for anthropological practice. Emerging themes from the experiences of AI specialists include the negotiability of technology, an ethics of the everyday and critical collaboration. Together this returns to an initial approach into a situated understanding of artificial intelligence, negotiating with broad, sensationalist perspectives and the more commonplace, backgrounded cases of narrow research.

show abstract

“…A growing number of researchers of algorithmic systems are making calls for the development of frameworks to understand data science as a situated practice, including reorienting their research questions "upwards," away from those who experience systematic disadvantage and toward those who inhabit positions of power and authority | Catalyst: Feminism, Theory, Technoscience Issue 6 (Vol 2) Sarah Myers West, 2020 (Elish & Boyd, 2018;Barabas, 2020;Katell et al, 2020). Others call upon technical designers to redraw their abstraction boundaries to include social actors, and not just technical actors (Selbst et al, 2018), and to embed a decolonial critical approach within technical practice (Mohamed et al, 2020). Such work is critical to overcoming the deep structural barriers that inhibit the development of a feminist "AI from below.…”

Section: Rekognition and Remediationmentioning

confidence: 99%

Redistribution and Rekognition

West

2020

Catalyst

View full text Add to dashboard Cite

Computer scientists, and artificial intelligence researchers in particular, have a predisposition for adopting precise, fixed definitions to serve as classifiers (Agre, 1997; Broussard, 2018). But classification is an enactment of power; it orders human interaction in ways that produce advantage or suffering (Bowker & Star, 1999). In so doing, it obscures the messiness of human life, masking the work of the people involved in training machine learning systems, and hiding the uneven distribution of its impacts on communities (Taylor, 2018; Gray, 2019; Roberts, 2019). Feminist scholars, and particularly feminist scholars of color, have made powerful critiques of the ways in which artificial intelligence systems formalize, classify, and amplify historical forms of discrimination and act to reify and amplify existing forms of social inequality (Eubanks, 2017; Benjamin, 2019; Noble, 2018). In response, the machine learning community has begun to address claims of algorithmic bias under the rubric of fairness, accountability, and transparency. But in doing so, it has largely dealt with these issues in familiar terms, using statistical methods aimed at achieving parity and deploying fairness ‘toolkits’. Yet actually existing inequality is reflected and amplified in algorithmic systems in ways that exceed the capacity of statistical methods alone. This article outlines a feminist critique of extant methods of dealing with algorithmic discrimination. I outline the ways in which gender discrimination and erasure are built into the field of AI at a foundational level; the product of a community that largely represents a small, privileged, and male segment of the global population (Author, 2019). In so doing, I illustrate how a situated mode of inquiry enables us to more closely examine a feedback loop between discriminatory workplaces and discriminatory systems.

show abstract

Decolonial AI: Decolonial Theory as Sociotechnical Foresight in Artificial Intelligence

Cited by 271 publications

References 103 publications

Artificial Intelligence, Values, and Alignment

Artificial Intelligence, Values, and Alignment

Coproduction, Ethics and Artificial Intelligence

Redistribution and Rekognition

Contact Info

Product

Resources

About