Methods and Critical Issues Training: History, Policy, Designing AI

Friday 24 July 2020, 15:00–17:00 BST

Overview

In 1869, J. R. Seeley wrote, 'History is the school of statesmanship.' How might historians today understand themselves as political actors versus seemingly neutral observers or interpreters? How does the craft pursuit of writing history relate to power structures engaged in the development of AI? This panel connects leaders who engage with historical methods in their own work, be that in the world of activism, design or policy, and asks them to share their views from outside academia.

Panelist: Sarah Hamid (Policing Tech Campaign Lead, Carceral Tech Resistance Network; Co-creator of 8 to Abolition)
Panelist: Dr. Dilan Mahendran (Independent Researcher; Google)
Panelist: Maikki Sipinen (AI Policy, European Commission)

Summary of event

In this session, panellists addressed topics concerning international AI policy, technological understandings of the human and the mind, and carceral technologies, resistance, and abolition.

Sarah Hamid's presentation related to her work and experience with the Carceral Tech Resistance Network (CTRN), the importance of knowledge-sharing, and scrutiny of sovereign databases used to aggregate information regarding migrants and BIPOC. She noted similarities in the work of historians and CTRN in relation to the analysis of the textual and cognitive matter of artefacts, identifying three important forms of work here: boundary work, historical critique, and epistemic intervention.

Maikki Sipinen's presentation addressed her work leading AI policy for the European Commission. She covered EU member states' documents, investments and commissions related to AI strategy. Focusing on recent years, she gave overviews of the public consultation on the AI white paper and the 2018 establishment of a high-level expert group to give advice on AI in relation to ethics and policy. She noted that multidisciplinary researchers had roles in this expert group, and that the role of policy makers can be helpfully understood through the work of historians.

Dilan Mahendran's presentation approached a wider epistemic question of what it means to be a human, rational subject. He delineated the ways in which Cartesian logics of mind/body divide have mapped onto our reasoning and presuppositions of what a computer or machine is, positioning this as a delimitation that can curtails the potentials of learning systems. He noted the importance of academic frameworks in relation to their translations into data.