In this webinar, Professor Shireen Kanji questions the neutrality of data as utilised in machine learning systems. AI systems are marketed on the basis that they can uncover objective truths by virtue of processing vast amounts of data. asks if we are justified in using data as the driving force in decision making and, if so, under what circumstances. It is certainly true that digital transformations have enabled the collection of vast quantities of data. But is it correct that adding more data resolves issues of accuracy, representativeness or understanding? Does big data do away with the need for theories to explain why phenomena exist?
Prof Kanji argues that AI is reconfiguring inequalities and that to understand both contemporary and emerging inequalities, we need to apply existing understandings from social science and history. She examines examples of what constitutes data, focusing on how it is collected and classified, looking particularly at data on social phenomena that have been used in algorithmic decision making. Investigating all stages in the chain reveals at what stage and how inequalities are created: from the nature of the data collected, how it is processed to what it is used for. Using retrospective data from unequal societies to predict the future also creates that future. Not only entrenching existing inequalities, but also limiting the possibility of breaking out of hierarchically lower social positions.
In terms of data quality, many social scientists would critique the poor quality of data gathered to make predictions which are highly consequential. For example, an algorithm to aid the US judiciary in decision making generates risk profiles using answers to questions such as “a hungry person has a right to steal”. New techniques of processing data are redrawing the ways in which inequalities emerge and persist, and how exclusion operates. For example, data can be used for digital red-lining, the practice of excluding certain groups of people. Facebook’s discriminatory ad-blocking and ad targeting tools have been shown to limit job vacancies advertised to women. Targeting potential renters by zip code reinforces racial segregation. Thus, these techniques are actively creating our collective, as well as individual, futures.
The AACCB webinars will help attendees to gain greater clarity about some contemporary issues, and how they link in with other related management research areas. Delivered by senior academics engaged in cutting edge research, the topics to be covered in the series will provide attendees with scholarly insight into what constitutes meaningful scholarship, and the making of optimal theoretical and methodological choices when crafting high impact manuscripts for publication. There will also be opportunities to discuss ways to get the best out of the publication process, crafting a careers, and networking with fellow academics, and other management researchers, all in one place.
BAM Council's Sub-Committee of Academic Affairs of Conference and Capacity Building (AACCB)
Who Should Attend?
The event speaks to Sections A1 and A2 as detailed in the BAM Framework
Benefits of attending
- Learn about new frontiers in management research methods
- Gain greater clarity on your methodological choices
- Network with other doctoral students and academics all in one place
Please contact the BAM Office at [email protected] with any queries.
Registrations close on Tuesday 5th July 2022 at 17:00 UK time.