Building a Better World: The Role of AI Ethics and Social Responsibility

Grewal, Guha, and Becker (2024) provide a compelling argument about how the profound impacts of artificial intelligence (AI) will continue to evolve over the next several decades. They do an excellent job of recognizing the potential benefits as well as the negative side effects caused by the rapid adoption of AI. Benefits include the ability to augment and improve human intelligence and emerging as an empathetic and trusted companion, but these benefits are challenged by novel tensions. These tensions include many ethical challenges such as bias, discrimination, job loss, or even weakened institutions. The social and economic fabric of our society could become unraveled with unknown outcomes. The objective of this commentary is to address and expand upon what Grewal, Guha, and Becker term ‘Theme #3-AI creates novel tensions.’ Specifically, we focus on AI ethics and social responsibility solutions.

Grewal, Guha, and Becker (2024) outline the efficiencies linked to AI for all marketing activities, including marketing strategy and implementation. Currently, AI can best augment rather than replace how marketing is performed. In the future, many marketing functions and tasks may be replaced as AI is able to learn, understand, and execute complex inputs (Guha et al. 2021). All aspects of the interface with customers can be changed by replacing human efforts and, as AI advances, the ability not only implement but also think and feel in marketing contexts (Huang and Rust 2021). In addition, marketing research will be able to use AI for segmentation, targeting, and positioning as well as the development of strategic marketing plans (Huang and Rust 2021).

Academic marketing research will advance, and traditional research will be enhanced and disrupted by AI. Consumer behavior research will use AI for data collection, experiments, and systematic literature reviews. AI will impact almost all areas of marketing, cutting across the 4 Ps of the marketing mix (i.e., product, price, place or distribution, and promotion) and affecting how impacts society. The novel tensions that will impact society include ethical challenges and how firms maintain a socially responsible focus. These challenges have not been adequately addressed by firms using AI, government regulations, or academic researchers. Therefore, this commentary will focus on how marketing ethics and social responsibility can be forces to ease the challenges and tensions related to the rapid adoption of AI in marketing. Our objective is to provide guidance for better understanding the challenges and opportunities for firms to maximize the benefits of AI and minimize negative consequences and harm that could result from using AI.

Ethical and Social Responsibility Challenges

The importance of ethics and social responsibility has become an important element of the marketing strategic planning process. A firm can see its image, reputation, and effective marketing activities destroyed by poor ethics and social responsibility efforts. Some believe ethics and social responsibility are good supplements and may not be essential elements of a marketing strategy. Yet, there are many studies that show ethics is directly correlated with many organizational performance elements including the bottom line (Ferrell, Fraedrich, and Ferrell 2022). Ignoring stakeholders’ demands for responsible marketing can destroy trust and can result in increased government regulation. In addition, when customers are harmed by irresponsible marketing activities, civil lawsuits can cause financial and reputational damage to the organization. If AI applications are not included in a firm's overall ethics and compliance programs, there can be significant negative consequences. Ethics and social responsibility are distinct constructs with different meanings but are linked by conceptual and operational relationships (Torelli 2021). Confusion is compounded by how these terms are used interchangeably by academics and the mass media. In most firms’ ethics and social responsibility are separated by boundaries in developing ethical and social responsibility initiatives (Weller 2020). These are different areas with few informal relationships. We will define and relate each concept to tensions in AI decision making.

Marketing Ethics

The term ethics relates to an individual's morals and values that can be applied to all decisions in life. Business ethics is different in that it relates to organizational principles, values, and norms that help create an ethical culture. These elements of the ethical culture can originate from individuals, organizational codes of ethics, or the legal system (Ferrell, Fraedrich, and Ferrell 2022). Marketing ethics is a subset of business ethics, related to the standards that guide individual and group behavior in making marketing decisions. Some of these standards have been codified into laws or oversight by regulatory agencies such as the Federal Trade Commission (FTC). Marketing ethics goes beyond platitudes with moral contexts such as honesty, respect, non-maleficence, trust, or integrity. It requires understanding the risk of decisions and the impact on various stakeholders. Marketing ethics requires communication and training of key decision makers to maintain shared values in ethical decisions.

Since AI is reshaping marketing strategy and implementation, understanding how ethical challenges should be managed is important. Most approaches to addressing ethical challenges in AI rely upon developing general principles to guide AI ethics without any directions on how to implement these principles in an organization. Microsoft, for example, adopted the principles of fairness, reliability and safety, privacy and security, inclusiveness, transparency, and accountability to guide the design of its AI systems (Microsoft 2024). Hermann (2021) examined published papers on AI and found principles under the headings of beneficence, non-maleficence, autonomy, and explicability injustice. These were found to be the most prevalent or practiced approaches to AI ethics. However, there are no directions as to how marketing managers communicating with AI developers can implement these principles. Principles continue to be the most discussed approach to AI ethics by firms, regulatory entities, and nonprofits (Ferrell et al. 2024).

Principles are normative and based on deontological foundations from philosophy. Principles could be considered a basic truth or duty. Unlike values, which are core beliefs selected or endorsed by an organization, principles are more permanent and enduring and should be applied across all organizations using AI applications (Ferrell et al. 2024). While there is general agreement that principles should guide AI ethical decision making, there have been few empirical studies to investigate how principals can be integrated into decision making (Kelley 2022) The limited coverage in the literature and lack of empirical investigation of using principles in AI ethical decision making suggests an important gap in creating ethical boundaries for AI decisions (Mustak et al. 2021) This gap exists because AI ethical decision making occurs in an organizational context with managers and application developers who may not be aware of ethical principles. Even if they know the principles, they must convert them into algorithmic rules to guide AI applications in performing operations. Rules perform computations that involve many steps and variables (Anayat and Gowhar 2024). Rules create formulas, blueprints, and directions that would include instructions for ethical AI outcomes. AI behavior relates to how humans program system applications with algorithmic inputs of data in training to make decisions (Huang and Rust 2021).

In discussing AI ethics, the focus needs to be on the human side of decision making and how these decisions are made. AI ethics seems more concerned with what are ethical challenges rather than how humans can resolve those challenges. Most developers have technical backgrounds in computer science, data science, software engineering, mathematics, and statistics (Baker-Brunnbauer 2021) Developers with these backgrounds need to rely on system managers for algorithmic rules for operations. If AI applications can develop to the level of thinking and feeling, the challenge is to have a system that can develop intentions to make an ethical decision.

The Theory of Planned Behavior links attitudes and behavioral controls together with subjective norms to perform a specific behavior (Ajzen 1991). Subjective norms relate to in-group significant others that can influence the final decisions that are made. Subjective norms have been found to be key variables in ethical decisions in an organization (Ferrell and Gresham 1985). Subjective norms cannot be programmed into AI applications currently but, as thinking and feeling develop in applications, it may be possible in the future. This would allow AI systems to not just depend on structured AI rule-based algorithms but embrace group or team influence in making the final decision. Without intentions in the ethical decision-making process, there should be human control from a team of participants injecting subjective norms as the AI application is developed and the evaluation of outcomes is determined. Davidovic (2022) proposes five common purposes for meaningful human control over AI and automated decision-systems, each requiring different institutional design: (1) safety and precision; (2) responsibility and accountability; (3) morality and dignity (4) democratic engagement and consent; and (5) institutional stability. Human controls should be included in each stage of the AI development process.

AI principles such as fairness, accountability, transparency, privacy, and explainability should be a focus at different stages of AI model development and deployment (Eitel-Porter 2021). A study by Behera et al. (2021) found that principles such as purpose, disclosure, and governance are critical to improving organizational effectiveness and reputation. A team effort to apply principles with the ability to discuss, contest, and exchange opinions is a critical safeguard to ethical decision making at the individual level (Lyons, Velloso, and Miller 2021).

A study of managers overseeing AI applications found that the implementation of AI ethical principles is lacking in practice (Harrison et al. 2023). While there is agreement in the use of principles managers do not have systematic approaches to guide developers. It was also found that developers rely more on subjective norms or significant others than their own confidence in making ethical decisions about an application (Harrison et al. 2023). This indicates the difficulty in managing AI application ethical outcomes. If there is a collaborative effort between managers and developers there can be improvement in translating principles into rules. There is a need for the organizational ethics and compliance function to be involved in the communication and training of AI managers and developers.

The entire organization needs to be involved in identifying the risks and intended consequences of AI. The level of effort to identify and control key risks goes beyond what most firms are engaged in today (Cheatham and Kia 2019). Identifying AI risks is difficult because the consequences of many applications are not known or involve disastrous repercussions, including the loss of life, and are not known or experienced yet. Since AI is being advanced at an increasing rate, managers have not had time to understand the full scope of societal, organizational, and individual risks (Cheatham and Kia 2019). In one example, Google's AI search feature provided users with information, including a suggestion to put non-toxic glue on pizza to help the cheese stick (Robison 2024). Although Google claimed these errors were isolated incidents tied to uncommon queries, the potential for misinformation is concerning, especially given the high level of trust people place in Google. Next, we examine AI's impact on society. Macromarketing focuses on how society is influenced by marketing. Macro marketing issues include social responsibility and the role of marketing in forming public policy. AI has the potential to have a major interface with these issues.

Social Responsibility Challenges

Social responsibility is an organization's duty to maximize its positive impact on stakeholders while minimizing its negative impact (Ferrell, Fraedrich, and Ferrell 2022). Social responsibility is a contract with stakeholders to make a positive impact. It takes an external perspective while marketing ethics takes an internal perspective (Bocean et al. 2022). Marketing ethics is the foundation of all decisions that could be judged as right or wrong and would include decisions about social responsibility. While all AI decisions have an ethical dimension our focus has been on the ethical decisions of those managers and developers making decisions about rules in AI applications. In other words, our discussion of AI ethics has focused on the ethical decision-making process that is done internally in the organization. AI application performance can have an impact on external stakeholders that can be related to external social responsibility impact.

Social responsibility issues apply to many AI ethics decisions related to social and economic issues impacting consumers, employees, institutions, and the health and welfare of society. Social responsibility overlaps with sustainability promoting the well-being of the natural environment, including all biological entities and mutually beneficial interactions between nature and all members of society (Ferrell, Fraedrich, and Ferrell 2022). AI applications can assist in climate change predictions, energy optimization, agricultural optimization, and especially healthcare accuracy, accessibility, and procedures. On the other hand, biased and inaccurate outputs on these issues could be negative to society.

AI can also relate to the risks of job loss, privacy issues, bias, and discrimination (Davenport et al. 2020). The challenge is to allow society to reap the benefits of AI while minimizing the negative outcomes (Grewal, Guha, and Becker 2024). At this point in time, the answer to how to achieve this goal is being debated by various stakeholders. There are many varying predictions on job loss. The World Economic Forum predicts 83 million jobs could vanish over the next five years (World Economic Forum 2023b). Marketing jobs that may be replaced relate to retailing, advertising, sales, pricing, and supply chain management. On the other hand, marketing managers that develop and implement strategy may become more important. Managers that oversee AI development and maintain the service of AI applications will increase. Applications that engage consumers with meaningful relationships in providing products will eliminate the need for some frontline service employees (Noble and Mende 2023).

While eliminating bias and discrimination in AI can be positive, the social impact of widespread adoption of AI may create a negative societal impact. Privacy ranks high on the list of negative impacts on society. Applications use large data sets with personal information that can be exploited through data leaks. Identity security will continue to be a constant threat to consumers. Cyber security will need to advance as criminals use AI to attack data sets. Algorithmic bias is possible as AI is trained on historical data leading to discrimination in all aspects of life. There are also dangers of socioeconomic inequality with unequal access to leaving marginalized groups disadvantaged. For example, Optum, a healthcare technology company, developed an algorithm to allocate care to the sickest patients, but the system was found to be racially biased as it recommended less care for Black patients compared to white patients with similar health needs. This bias arose because the algorithm relied on healthcare expenditure as a proxy for illness, which overlooked systemic disparities in healthcare access and spending (Weber 2023).

As AI systems advance there could be catastrophic world events from weapon automation. Wars could be fought with AI-equipped systems taking actions without human direct control with weaponized robots, drones, aircraft, tanks, and submarines being employed. Autonomous weapons could make serious errors in identifying targets through malfunctions and misinformation. Already cheap drones are being used to swarm targets without adequate information of who or what may be destroyed. AI systems could outthink and process information much faster than humans. This AI ability could lead rogue nations to inflict much damage on another country at will.

The legal and regulatory community has not been able to catch up with the capability of AI to disrupt the economic and social system. Major institutions that establish rules and norms to shape and constrain society, such as education and government, have many challenges in understanding how AI is being used to accomplish tasks. Generative AI has disrupted education, and schools and universities are having to adjust rapidly to solve conflicts. While some educational institutions banned the use of ChatGPT to prevent cheating, now the concern is how to integrate ChatGPT by establishing ground rules and guidelines for its appropriate use (World Economic Forum 2023a). For example, the University of North Carolina quickly recognized that banning generative AI tools was impractical and could disadvantage students, so they established the UNC School of Data Science and Society, a committee consisting of faculty and staff from across the university, to create guidelines for using generative AI in instructional settings (Dean 2023).

There is no comprehensive federal legislation on AI in the United States, but there is general legislation that covers harm to consumers from negative outcomes associated with AI. Many states have introduced bills for legislation on AI. The United States Congress has asked the FTC to recommend reasonable policies and procedures to address harm to consumers. There has been an executive order on safe, secure, and trustworthy AI. The order supports testing standards and public-private collaborations. Government departments and agencies are to develop frameworks for responsibilities but national security is the main focus (Ferrell et al. 2024).

The most comprehensive AI legislation is the European Artificial Intelligence Act which went into effect in 2024 across all 27 EU states. It provides risk-based categories for AI applications. The classification system ranges from minimal risk to applications that are banned. Unacceptable risk or manipulative systems potentially lead to harm and are outlawed. High-risk applications include those used for evaluating creditworthiness, educational institutions, and infrastructure. Limited risk relates to user interaction with AI devices such as reminder systems and chatbots. Minimal risk related to applications such as spam filtering, video games, and any other area not of high risk. The Act provides compliance rules not directly translated into algorithms but rather a guide to algorithmic development (Regulation [EU] 2024/1689).

To address emerging laws and regulations, firms will need to incorporate AI risk management in their corporate governance as well as ethics and compliance programs. Corporate governance relates to the CEO, other officers, and the board of directors providing oversight of organizational decision-making and resource allocation. The board should approve a code of ethics that includes the use of AI applications. Managers of AI systems should be in collaboration with ethics and compliance officers to assess risks. There should be a transparent alignment between corporate ethics, AI managers, and developer oversight. At this point, developers might be creating applications without proper risk assessment and controls, leading to potential cases of bias and discrimination. Internal and external audits are an important element in managing risk, controls, and corporate governance (Ferrell, Thorne, and Ferrell 2024) Through risk management and a strong internal control system, AI decision makers should assess AI risk related to outcomes that could impact key stakeholders. As AI advances into new areas managers should develop systems for identifying how AI applications can have a positive or negative effect on consumers, employees, and society. Based on available research, AI issues are being identified as they occur, but there is limited evidence that AI managers and developers are included in overall ethics and compliance programs (Harrison et al. 2023). The biggest risk concern is the pace of AI development and understanding its capabilities. With regulatory oversight just emerging to guide AI development, the focus has been on advancing the technology rather than focusing on ethical frameworks to manage ethical decision making. Because the impact of AI on society is not known, Elon Musk, Steve Wozniak, and hundreds of industry leaders suggested a six-month pause on training and development of AI systems to create better safety protocols (Newcomb 2023). For example, academic research is being conducted at record levels, more than humans are capable of reading. AI's use in developing and evaluating knowledge may be corrupting the integrity of knowledge production. AI tools also may be a downside for intellectual exchange (Lee 2024). While this is not violating existing laws or regulations it is a major risk to academic research and to society. Many risks to society fall into this category. These types of issues may not be regulated but can have an impact on society.

While there is much documentation of AI issues such as discrimination and manipulation in the media, it is important for AI managers and developers to report near misses and create a near miss incident database (Shrishak 2023). Systems have harmed the basic rights of people, such as housing discrimination, and tracking these harms as incidents would help AI system developers and government regulators to hopefully avoid these incidents in the future (Shrishak 2023). This is a good example of operationalization to bridge normative ethical principles to how AI risks exist in practice. There has been a theory gap in trying to relate principles to address actual risks that exist (Bleher and Braun 2023). While principles can be a normative background, embedding ethics and social responsibility knowledge about risk into the processes and activities of managers and developers could help improve decision making. In other words, the technical education of developers should be integrated with ethics and social responsibility knowledge (Bleher and Braun 2023).

To accomplish operationalizing of principles in the education and training of all participants in AI decision making will require using the existing infrastructure in organizations. For example, using existing organizational functions such as risk management, internal control, ethics and compliance, finance, marketing, and human resources awareness and feedback to support AI ethics and social responsibility can be achieved. By building organizational awareness and creating a culture of recognition, AI ethical considerations as a part of all decision making could be achieved (Blackman 2020).

Discussion

While this commentary focuses on ‘Theme #3-AI creates novel tensions’ described by Grewal, Guha, and Becker (2024), the authors provide many benefits in Stage 1: Enhancing firm efficiencies and Stage 2: Bringing the individual to the forefront before addressing the concerns in Stage 3: Rising societal concerns. They point out that the boundary between Stage 2 and Stage 3 remains undefined but predict that Stage 3 will start when the negative impact of AI begins to outweigh its positive impact (Grewal, Guha, and Becker 2024). As AI advances to Stages 1 and 2 it will be able to exceed humans in thinking and feeling and become social AI where it exhibits empathy and builds long term, lasting meaningful relationships with humans (Noble and Mende 2023). In all stages, there will be ethical and social responsibility challenges. What will be lost in humans if they lose their creativity and innovation capabilities (Grewal, Guha, and Becker 2024)?

Our approach has been to focus on the ethics and social responsibility issues that exist today, but there are far reaching questions about the future impact on individuals and society. AI could create new forms of creativity by offering novel perspectives and insights inspired by humans. AI could actually not only augment human creativity but also help humans become more creative through efficiency and productivity.

Over the last 100 years technology has increased at an exponential rate with many positive impacts on society. Automobiles, aviation, television, computers, the internet, and smartphones have all impacted society. They have contributed to access to travel and increased our ability to communicate globally. This social connectivity has led to economic growth and many advances in social issues like healthcare. While there are benefits, there are always negative impacts with privacy and security as possibly the biggest societal concerns. There is reduced personal interaction with social media, texting, and other forms of engagement linked to mental illness, harassment, misinformation, diminished social skills, isolation, anxiety, addiction, and unrealistic expectations about life (Brown 2018). On the positive side, social media has opened up the opportunity for marketing to gain cost-effective consumer engagement and brand awareness.

Computers are the foundation of most technology at this time. Computers, similar to AI, create efficiency and greater capability to handle many tasks. It has also facilitated social and economic development. While computers are a tool that can be used for negative impact on society, overall, computers have been more positive the negative in their impact and influence. AI builds on this foundation to expand how humans using technology can achieve progress for our economic and social well-being.

The history of marketing has been driven by innovations in technology. The automobile changed the landscape of marketing in almost all activities. Mass marketing and brand loyalty evolved with shopping centers, big box stores, and supply chains. Cars and trucks completely changed consumer behavior. Radio allowed mass marketing with real-time advertising having a profound impact on consumer behavior. Television created the demand for products as a powerful medium for mass communication with visual and audio combined. Television changed the norms and values affecting culture and all institutions. The introduction of the computer changed data analysis and created new marketing channels such as websites, e-mail, search engines, and other platforms to precisely target consumers. E-commerce has transformed retailing and expanded the reach of both consumers and even small businesses.

AI is the latest innovation to revolutionize and change marketing similar to the earlier innovations. It advances marketing decisions with predictive analytics, content creation, insights into consumer behavior, and customer service as well as many other marketing activities inviting concerns that the impact on consumers is different from other technologies. It can augment and (potentially) replace human intelligence and evolve into an empathetic and trusted companion (Grewal, Guha, and Becker 2024). Possibly one of the major issues will be personalization in all aspects of life. Daily interactions to facilitate shopping, healthcare, and education as well as most aspects of life will once again have an impact on social norms and cultural practices. Since AI is a relatively new technology that may have more impact on society than previous technologies, there are many uncertainties. When ChatGPT was asked “Will AI destroy the world?”, it responded, “It might be time to move on to a new topic.” This invites the question of whether AI will destroy human existence as we know it, a topic of significant debate among many experts.

Beyond ethics and social responsibility, the question could be: Can AI reach a point where it advances its ability to uncontrollable levels? A 2023 survey of AI experts found that 36% fear AI may result in a nuclear-level catastrophe (Hunt 2023). We take the position that while AI may hold potential extreme existential risks, this should be balanced with current risks such as bias, discrimination, and fairness. There are tremendous benefits, and we have not reached the point where negative outcomes are greater than positive impacts. While AI can be used as a tool for malicious purposes, it is humans that are in charge of using AI to that end.

Human designs, codes, programs, and training of AI applications are critical to successful implementation. Therefore, there is an opportunity for human controls at each stage of the development and implementation process. AI ethical decision-making models emphasize the human control variable at each stage of AI development (Ferrell et al. 2024). AI development should be under the control of AI managers and developers who understand how ethical principles can be translated into rules and algorithms. Humans are necessary to program accountability and responsibility and determine how machine learning incorporates ethical and socially responsible decisions. Allowing managers and developers to only focus on efficiencies and accomplishing tasks opens the door for negative ethical outcomes. Many of the negative outcomes to society are hard to control with ethical decision models and regulations.

Dealing with the challenges of (1) preserving and growing human capability; (2) protecting societal belonging and human connection; and (3) ensuring the sharing of AI benefits will be important to achieve (Grewal, Guha, and Becker 2024). Preserving and growing the human capacity will require leveraging AI to promote productivity and efficiency. Advancing education, skill development, health care, and sustainable practices through AI will address this challenge. These societal outcomes can evolve through responsible AI, protecting societal belonging and human connection. These outcomes will require AI to be used to enhance face-to-face human interactions. AI can help build a culture of connection to combat social isolation and feelings of loneliness. Sharing AI benefits will require the adoption of ethical frameworks that avoid bias and discrimination and provide transparency as well as accountability. Addressing these macro societal outcomes goes beyond the ethical AI decisions of one organization but will require institutional support from government, education, regulatory agencies, and world economic organizations. Military institutions will play a crucial role in the responsible use of AI. The United States Department of Defense released an AI adoption strategy for improving decision-making, enhancing battlefield effectiveness, and ensuring ethical AI applications (Clark 2023). The military is integrating AI into everything from business operations to delivering capabilities to the battlefield.

While the impact of AI goes beyond marketing-related AI applications, marketing must understand the benefits that AI offers as well as the perils it creates now and in the future (Grewal, Guha, and Becker 2024). Without controls, autonomous AI decision-making could evolve at an increasing rate, surpassing human intelligence. Marketing needs to implement control strategies to ensure there is consensus among consumers and society and that affected entities are served as well as protected. The goal should be to align the use of AI in marketing with human values and safety. Marketing needs to be aware of the novel tensions AI creates and be proactive in finding solutions.

We suggest that AI ethics and social responsibility be incorporated into the marketing strategic planning process. Government regulation is evaluated at a very slow pace with the regulatory community not able to keep up. Marketing needs to identify risks and develop standards that are communicated across the entire organization. AI has too many risks in the development of AI applications to be decided by those with only technical expertise. Ethical AI decisions should be based on principles that are permanent and enduring. The firm should develop shared values that are core beliefs to guide AI decisions in marketing. Limited evidence indicates that ethical principles have not been effectively integrated into AI application development (Harrison et al. 2023). To accomplish this, top leadership and governance systems need to focus on the implementation of this approach to AI ethics.

Responsible AI and marketing can minimize the negative impact on society. AI can be a positive or negative force in the welfare of many stakeholders. While consumers can benefit from better service and personal intelligence, loss of privacy, jobs, and alienation could create inequality. Social responsibility issues that evolve from AI are more from the collective use of AI and marketing, not just the decisions of one firm. These challenges may be best addressed by institutions including regulation, education, family, and socialization to maintain a core cultural foundation. There are many questions to answer as to how to use AI for a better world.

Conclusions

The purpose of this commentary is to extend the discussion of the theme that AI creates novel tensions (Grewal, Guha, and Becker 2024). Many of these tensions emerged through marketing-related AI applications. Tensions from risks associated with privacy, bias, discrimination, worker autonomy, and job loss are long-term societal challenges. These challenges go beyond the ability of marketing to control, preserve, and grow human capability, protecting the human connection and ensuring equitable benefits. This will require ethical AI development with transparency, accountability, and fairness in all AI systems. AI safety research, regulation as well as public awareness and education will be needed to change the world for the better.

For marketing, there is the opportunity to integrate AI ethics and social responsibility into marketing strategy. AI cannot be just a tool or tactic to ensure efficiencies, but marketing planning needs to identify risks as well as opportunities to responsibly serve consumers and other stakeholders. AI ethical decision making occurs in an organizational context. There's evidence that there is not always a seamless flow of communication between managers and developers in incorporating ethical principles into rules for algorithms. We suggest that AI ethical decision making can be improved by including the organizational ethics and compliance function in the communication and training of AI managers and developers. All marketing managers need to understand the opportunities for principal-based decisions that serve and protect consumers. While the future of AI cannot be predicted, marketing can take a leadership role in responsibly using AI.