AI induced reflexive environmental governance : examining artificial intelligence and related digital technologies' abilities to reduce environmental exposures to endocrine disrupting chemicals in the plastics and pesticides sectors

Abstract

Theoretical research regarding artificial intelligence (AI)’s role in global environmental governance is new and generally lacks holistic, empirical evidence that can build our understanding of AI within governance theory. Meanwhile, evidence of failures of current environmental regulatory regimes continues to build. This includes regimes that inadequately protect humans and the environment from the negative impacts of exposures to endocrine disrupting chemicals (EDCs), through their lack of appropriate identification criteria, testing methods and bias and fraud from the participation of vested industry actors. Systems Theorists explain environmental governance failures as being due to normative communication deficits within social system dynamics. Some believe the economic system has ‘hyper-socialised’ all other social sub-systems and see the economy’s deficiency of environmentally normative communication, combined with its expansionist tendencies towards ever increasing profits, as the cause of current environmental governance failures.

To ameliorate this normative deficit, Systems Theorists propose reflexive solutions, such as Teubner’s sociological interpretation of the theory of societal constitutionalism, Orts’ Reflexive Environmental Law and Meadows’ interventions in system ‘leverage points’. The plastic and pesticide sectors represent two of the largest sources of environmental EDC exposures and this thesis undertakes an empirical, desktop scoping study of AI and related digital technology applications of the 31 largest plastic and pesticide producers, consumers, financiers and investors. It analyses the applications’ abilities to reduce environmental EDC exposures, assesses whether they evidence reflexive solutions and examines the applications’ impacts from a broader, theoretical environmental governance perspective.

The scoping study identifies 89 applications that exemplify reflexive solutions with potential to reduce environmental EDC exposures. The applications also broadly reveal three unique communicative AI governance processes. First is their unequalled ability to reduce information gaps and asymmetries in the economy through the unmatched scale and detail of their data gathering and pattern identification capabilities. Second is their extraordinary ability to communicate environmental norms directly into the economy by ‘translating’ non-financial data into quantified economic ‘vocabulary’. This allows AI’s predictive and generative capabilities to reflexively and dynamically induce data-driven, environmentally normative, economic decision making that can circumvent incumbent power structures. Third is their novel ability to become the economy’s primary communicative medium (money and payment operations) and create permanent, environmentally normative counter structures against the economy’s expansionist tendencies.

It is proposed that cumulatively, with much caution to their risks and weaknesses, these new AI governance processes evidence an emerging AI induced reflexive environmental governance model. Expanding Systems Theory and advancing Cybernetics’ study of the control of the flow of communications in systems and the nascent theoretical concepts developing for AI environmental governance, the scoping study identifies the beginnings of these innovative, decentralised, data-driven, digital, transnational governance processes, which facilitate the potential for a dynamic, hybrid, algorithmic ‘mycorrhizal’ communication network of societal norms. These governance processes reflexively catalyse change from within the economic system by shifting the corporations in the scoping study’s normative beliefs regarding how they should function. In doing so, these applications suggest AI provides an unprecedented reflexive environmental governance model for future theoretical research and practical applications.