Corporations that are able to use Artificial Intelligence (AI) in the most suitable way are not only likely to overcome the current COVID-19 crisis more effectively, but also to succeed in the digital economy. AI can indeed play a key role in corporate boards, but it also creates significant risks, which can only be properly addressed if the corporate structure is designed to cope with a more extensive use of AI.
Our recent empirical study analyses the role of tech committees dealing with AI within European and North American listed companies to understand the level of AI governance in place.
The starting point for the creation of our dataset is WRDS-BoardEx, which features data about directors, those serving on a given committee in a defined geographic environment, on the one hand, and the specifics about individual directors sitting on the board, on the other hand. We considered listed companies in the period 2000-2019 to study the trends in the adoption of tech committees and focused on data from 2019 only to explore tech committees’ composition and functions performed. From 2000 to 2019, the number of companies that used a tech committee was 21.997, of which 5.313 were listed in the EU (excluding the UK) and 16.684 in North America (US and Canada). Instead, in 2019, only 28 European companies and 56 North American companies had a tech committee in place.
The study is carried out firstly from a quantitative point of view, also taking into account the personal characteristics of the relevant directors, and secondly from a qualitative standpoint, exploring specifically the functions performed by the committee at hand.
First of all, the paper shows that, in the 2000-2019 period, tech committees initially made their appearance overseas, later spreading to Europe. Although EU companies embraced this board structure, they did it in much smaller absolute terms and in slightly smaller percentage terms compared to all listed companies. At the same time, EU companies’ tech committees have a greater weight in terms of capitalization in the relevant national market. What is particularly interesting is how, in Europe, these tech committees are mainly established in banks—which presumably reflects the potential that these companies foresee in the development, also in the regulatory environment, of these systems and their growing use for monitoring purposes.
Secondly, as to the composition of tech committees in 2019, clichés such as the predominant involvement of independent directors are confirmed. In addition, tech committees unexpectedly feature directors who, on average, are more senior than non-tech ones (average age of directors serving in a tech committee is 74 in North America and 62 in Europe, while the average age of director serving in the other committees is 63 in North America and 60 in Europe). The latter point can be explained because, for the time being, such committees—as it emerged from the qualitative analysis—deal with strategic issues rather than with ‘new technologies’, as one would presume from the committees’ name.
The qualitative analysis of the functions actually undertaken by the tech committees identified above is aimed at understanding their precise role within the board of directors. To this end, a dataset comprising the committee charters (or, at least, the descriptions that can be retrieved from the company website to portray the activities of said committees) of the companies that offer this information has been developed. The hand-picked database of excerpts is then examined through content analysis, which reveals five main activities, ie ‘Strategy’, ‘Monitoring’, ‘Innovation’, ‘Risk Management’ and ‘Security’. Then, the same database was examined through a qualitative comparative analysis, performed using Stata, which allowed us to draw meaningful conclusions: We observed that tech committees are mainly engaged in the typical board duties, such as monitoring and strategy, while they also play a minor role in the exploitation of AI potentialities and, above all, in the management of AI-related risks. The two geographical contexts under review (continental Europe and North America) are characterised by the fact that tech committees do not seem to be ‘interested’ in understanding the functioning of AI systems, nor in exploring their capabilities with respect to specific corporate needs. Instead, tech committees perform more typically strategic tasks, especially where their core business is closely related to technological development, or monitoring on behalf of the board of directors.
The results of our two-level empirical analysis lead us to believe that the use of technology in corporate governance, in whatever form, now calls for an AI-oriented governance—and even more so it will in the near future— in particular by implementing shrewdness and caution when it comes to the characteristics and functions of the individuals within both committees and the boards.
We, therefore, propose a two-layer AI governance that corporate boards should adopt in order to fully benefit from the latest technological developments. The more operational level would thereby be informed and modelled according to the principles developed at the systemic level for the development of a reliable AI such as (i) human intervention and surveillance; (ii) technical and security robustness; (iii) confidentiality and governance of data; (iv) transparency; (v) diversity, non-discrimination and equity; (vi) social and environmental well-being; and, finally, (vi) accountability. Indeed, we believe that a sound tech AI governance cannot neglect the ethical implications of using such systems, and tech committees can develop as the venue where ‘virtuous’ AI usage can be planned, implemented and, now more than ever, governed.
Maria Lillà Montagnani is Associate Professor of Commercial Law at Bocconi University and Transatlantic Technology Law Forum Fellow at Stanford Law School.
Maria Lucia Passador is a post-doctoral researcher at the Université du Luxembourg, as well as an LL.M. Candidate and John M. Olin Fellow in Empirical Law and Finance at Harvard Law School.