Companies initially pursued the public interest over profits. However, today's companies tend to follow shareholders rather than considering other stakeholders. For example, a mining company developing a mining area may take necessary measures to continue its business to make profits for the economic benefit of shareholders, even if there are negative impacts on the environment and people around the mining area. This behavior is also seen in the artificial intelligence (AI) industry. The reluctance of some AI companies to circumvent regulatory compliance and governance indicates that the profit perspective still exists in this industry amid public concerns that AI is a potentially harmful technology. Therefore, a better governance concept is needed to guide AI companies towards more responsible enterprises.
Corporate Social Responsibility (CSR) is a governance concept that directs companies to operate responsibly while considering other stakeholders that may interact with the company. CSR helps support the spread of AI ethics and AI governance as it balances profits within a company with non-profit goals.
CSR begins when the law is silent
CSR is a long-standing concept that developed as a response to the realization of a better role for corporations in the community. Michael Blowfield and Alan Murray state that CSR starts where the law ends, i.e. CSR complements laws and regulations where these legal instruments fail to cover. This discussion completes the concept of a stakeholder societal perspective. Bob Tricker states that this theory moves away from stakeholder concerns to balance responsibility, accountability, and power in society. Thus, we see two elements of CSR that are relevant to AI governance:
First, CSR is a flexible yet actionable concept. Amid tensions between companies and governments over whether to regulate AI business activities, CSR may answer regulators' governance dilemmas. It is understandable that companies, especially AI companies, tend to prioritize resources on product research and profitability. To date, AI is an evolving technology that has been hindered by negative perceptions of safety and reliability. Therefore, the application of CSR concepts needs to be aligned with a company's corporate governance, including risk management practices.
Secondly, CSR acts as a “carrot” and a “stick” with respect to the regulatory regime. The “carrot” is that the application of CSR allows companies to enhance their reputation as ethical and trustworthy business actors. As a result, companies can enjoy financial benefits, such as easier access to bank loans and investments from institutional investors. The “stick” is that companies apply CSR as a regulatory obligation. Indonesia applies this by requiring companies to implement dedicated and planned CSR within their annual budgets and activities. Therefore, there is an element of sanctions for failure to comply with this obligation, which is stated in PP No 47 of 2012, and the sanctions shall refer to existing laws and regulations.
CSR supports AI ethics
From the above, we can see that CSR supports the application of law. To some extent, CSR can also support the dissemination and implementation of AI ethics. According to UNESCO's Recommendation on AI Ethics, AI ethics shall be achieved through the implementation of government-led ethical governance. Therefore, we can conclude that AI ethics is developed under a proper governance system that is applied by all parties, including AI companies. CSR can support AI ethics by making ethical AI a corporate value. In this way, the value is implemented by policies and governance. Governments can choose whether to make ethical AI governance a “carrot” or a “stick” goal.
Challenge
Although this idea is interesting, there are challenges to be addressed. CSR standards are not yet unified. As a result, companies may choose the scope and areas of their business that they want to manage with CSR. Therefore, AI ethics as part of CSR has some inaccuracies that are perceived as hiding jargon and policies without actual implementation. Therefore, it is important to create policies to standardize the application of CSR-AI ethics. Governments can enact laws or promote international standards. By establishing government initiatives, AI ethics will be easier to implement and can have a greater impact on society.
