A sociotechnical perspective for the future of AI: narratives, inequalities, and human control

Abstract Different people have different perceptions about artificial intelligence (AI). It is extremely important to bring together all the alternative frames of thinking—from the various communities of developers, researchers, business leaders, policymakers, and citizens—to properly start acknowledging AI. This article highlights the ‘fruitful collaboration’ that sociology and AI could develop in both social and technical terms. We discuss how biases and unfairness are among the major challenges to be addressed in such a sociotechnical perspective. First, as intelligent machines reveal their nature of ‘magnifying glasses’ in the automation of existing inequalities, we show how the AI technical community is calling for transparency and explainability, accountability and contestability. Not to be considered as panaceas, they all contribute to ensuring human control in novel practices that include requirement, design and development methodologies for a fairer AI. Second, we elaborate on the mounting attention for technological narratives as technology is recognized as a social practice within a specific institutional context.

Introduction. Artificial intelligence (AI) is not a new field, it has just reached a new ‘spring’ after one of the many ‘winters’ (Boden, 2016; Floridi, 2020). As a matter of fact, we might be on the brink of a new winter since different actors (firms, individuals, media and institutions) have concretely started questioning the over-inflated expectations. It may be the multiple ongoing narratives, including the ones of moving from the traditional ‘black-box approach’ to the use of transparent and explainable methods (Guidotti, 2019a, 2019b), the ‘scary’—but improbable—prospects of creating a ‘superintelligence’ that will convert humans into paperclips (Bostrom, 2014), or even of an ‘AI race’ between nations for the development of the ‘ultimate’ algorithm (Houser & Raymond, 2021). Then again, the term ‘AI’ means different things to different people; anything from data aggregation and manipulation to ‘magic’ (Theodorou & Dignum, 2020). Yet, AI is neither a myth nor a threat to man (Samuel, 1962).

Discussion / Conclusion. As we have been emphasising in our paper, AI practice is interdisciplinary by nature and it will benefit from a different take on technical interventions. Nor superior nor more appropriate, technical considerations (such as objectivity, fairness, and accuracy) should go in parallel with other types of knowledge useful for social change (Green, 2019). What issues to face, what data to use and what solutions to implement are compelling, not old-fashioned, questions. It is not always a question of efficiency and accuracy, but also it is about inclusivity by bridging the gap between technical and social research in AI. In addition to responsibility, AI should be inclusive, built upon quality data that comprises gender, education, ethnicity, and all of the other social and economic differences that are sometimes determining factors for inequality. Quality data not only means to make it respectful of privacy but to make it inclusive when it comes to social concerns and purposes.