Risks and opportunities of the Artificial Intelligence wave
15 AUG, 2024
By Deepshikha Singh
By: Deepshika Singh, Head of Stewardship, Crédit Mutuel Asset Management
Artificial Intelligence (AI) and digitisation carry significant societal risks, from job losses in traditional industries to increased safety concerns and rising discrimination. Generative AI (GenAI) could automate a significant part of a job's tasks, leading to potential job losses in the most affected occupations. There is a great risk of misalignment between the short-term financial incentives of AI development and deployment, specifically with GenAI, and the interests of humanity. A recent study by Capgemini Research Institute revealed that 72% of consumers are concerned about the misuse of GenAI technology. According to the OECD AI Policy Observatory's AI Incident Monitor (AIM), examples of these risks have massively increased since the beginning of 2023.
There are also negative environmental impacts related to GenAI throughout the value chain that need to be taken into account. Carbon emissions are expected to increase throughout the value chain. Competition to build data centre infrastructure has also raised questions about the ability of national energy grids to cope with the expected increase in electricity demand linked to AI, and whether there is sufficient renewable energy generation in these markets to power the technology. E-waste and the need for rare minerals and metals for infrastructure and production of GenAI applications are other potential risks we need to consider.
Investors cannot address the multifaceted risks posed by the rapid adoption of AI and GenAI in the last two years all at once. Engagement and management tools will be most effective. Investors need companies (both developers and implementers) to implement responsible AI practices throughout the organisation to protect against the social and environmental risks posed by AI. Organisations need to establish clear principles on how to apply GenAI and establish guardrails that ensure its safe application and specifically avoid bias, discrimination, misinformation and privacy violations. While some environmental impacts, such as end-user energy consumption and data centre energy efficiency, may change with the overall decarbonisation of the grid, investors are increasingly concerned that the technology sector meets its climate targets. Both AI developers and deployers will need to invest substantially in creating additional renewable energy.
Regulation can also help. The EU's AI Law, formally adopted last month, is a good risk-based framework for assessing and analysing a company's responsible AI governance and risk management systems. In addition, several engagement toolkits have been published in the past six months, such as the World Economic Forum's Responsible AI Strategy Book for Investors (June 2024), the Australian Investors Association's (RIA) AI and Human Rights Investor Toolkit and AI: An engagement Guide by International Corporate Governance Network (ICGN) (March 2024), which can guide investors in creating a robust engagement framework for responsible AI.
In the asset management industry, AI models can be used to create innovative quantitative investment strategies and risk management processes, as well as to improve portfolio performance. According to a Harvard study, GenAI can help speed up financial analysis, such as reading thousands of pages of data files to analyse a company's earnings and future trajectory, and analysing huge data sets to detect insights that humans simply would not be able to detect or do not have time to detect. La Francaise Systematic Asset Management, the asset management company of Groupe La Française (holding company of Credit Mutuel Alliance Fédérale's asset management business line) has created a system to integrate complex non-linear relationships and interactions in financial data. The system uses state-of-the-art machine learning models and combines them with a proven traditional behavioural model to detect endogenous shocks early, responding dynamically to the changing market environment. The same principles can be applied to sustainability analysis, where AI tools have been deployed to capture risks more quickly and effectively. AI-based tools have proven to be very effective in capturing controversies and governance failures from anywhere in the world that are impossible for human analysts to track.
