Fundamentals

Consider this ● a local bakery, beloved for its sourdough, decides to implement an AI-powered inventory system. Sounds innocuous, perhaps even smart. Yet, beneath the surface of streamlined efficiency lies a complex web of moral considerations. It is not merely about algorithms and automation; it is about how these technologies reshape the very fabric of small and medium-sized businesses (SMBs), the cornerstones of our communities.

The Unseen Algorithmic Hand in Small Business

For years, SMBs thrived on personal connections, intuitive decision-making, and a deep understanding of their local markets. Artificial intelligence Meaning ● AI empowers SMBs to augment capabilities, automate operations, and gain strategic foresight for sustainable growth. promises to augment, even replace, some of these human elements. Imagine an AI tool analyzing customer data to personalize marketing emails for a boutique clothing store. This seems like a smart move, right?

Perhaps. But what if the AI, trained on historical sales data, inadvertently reinforces biases, targeting only affluent customers and neglecting lower-income segments of the community? The moral implications begin to surface.

The integration of AI into SMB operations Meaning ● SMB Operations represent the coordinated activities driving efficiency and scalability within small to medium-sized businesses. introduces a new layer of complexity to everyday business decisions. It is no longer solely about human judgment; it is about entrusting algorithms with tasks that directly impact customers, employees, and the broader community. This shift necessitates a critical examination of the moral dimensions embedded within these technologies.

Fairness and Bias ● Are AI Systems Inherently Neutral?

The myth of algorithmic neutrality is pervasive. Many assume that because AI is based on code and data, it is objective and unbiased. This assumption is dangerously flawed. AI systems are trained on data, and data reflects the biases of the world it represents.

If historical data used to train an AI hiring tool for a small accounting firm predominantly features male candidates in leadership roles, the AI might inadvertently perpetuate gender bias, favoring male applicants over equally qualified female candidates. This outcome is not a deliberate act of malice; it is a consequence of biased data and algorithms that learn to replicate existing patterns, regardless of their ethical implications.

AI in SMBs is not just a technological upgrade; it’s a moral reckoning.

For SMBs, the implications of biased AI can be particularly acute. Limited resources may prevent thorough auditing of AI systems for fairness and bias. A small restaurant using AI-powered customer service chatbots Meaning ● Customer Service Chatbots, within the context of SMB operations, denote automated software applications deployed to engage customers via text or voice interfaces, streamlining support interactions. might unknowingly deploy a system that is less responsive or helpful to customers from certain demographic groups, damaging their reputation and alienating valuable clientele. The challenge for SMBs is to navigate the allure of AI efficiency while remaining vigilant about its potential for perpetuating and amplifying societal biases.

Transparency and Explainability ● The Black Box Problem

Many AI systems, particularly those based on complex machine learning Meaning ● Machine Learning (ML), in the context of Small and Medium-sized Businesses (SMBs), represents a suite of algorithms that enable computer systems to learn from data without explicit programming, driving automation and enhancing decision-making. models, operate as “black boxes.” This means that even the developers who create these systems often struggle to fully understand how they arrive at specific decisions. For an SMB owner, this lack of transparency can be unsettling. Consider a local hardware store using an AI-powered pricing tool to optimize profit margins.

If the AI suddenly recommends a significant price increase on a particular item, the owner might be left wondering why. Without explainability, it becomes difficult to assess the rationale behind AI-driven decisions and to ensure they align with the business’s ethical values.

Transparency is not merely about understanding the technical workings of AI; it is about maintaining accountability and trust. If an AI system makes an error that harms a customer or employee, the lack of transparency can hinder the ability to identify the root cause and rectify the situation. For SMBs, which often rely on strong customer relationships and community trust, the black box nature of some AI systems poses a significant moral challenge. How can a business be ethically responsible for decisions made by a system it does not fully comprehend?

Accountability and Responsibility ● Who is to Blame When AI Goes Wrong?

As AI systems become more integrated into SMB operations, the question of accountability becomes increasingly complex. If an AI-powered system makes a discriminatory hiring decision, who is responsible? Is it the SMB owner who deployed the system? Is it the AI vendor who developed it?

Or is it the algorithm itself? The lines of responsibility blur in the age of AI. For SMBs, this ambiguity presents a significant moral and legal risk.

Imagine a small e-commerce business using AI to manage customer orders and shipping. If the AI system malfunctions and misdirects a large number of orders, causing significant customer dissatisfaction and financial losses, who bears the responsibility? The SMB owner is ultimately accountable for the business’s operations, but they may lack the technical expertise to fully control or understand the AI system. The vendor may argue that the malfunction was due to unforeseen data inputs or user error.

The algorithm, of course, cannot be held responsible in any meaningful sense. Establishing clear lines of accountability is essential for ensuring that AI is used ethically and responsibly in SMBs.

The Human Element ● Maintaining Values in an Automated World

SMBs are often characterized by their human touch. They are places where customers are greeted by name, where employees are treated like family, and where business decisions are guided by personal values. The integration of AI risks eroding this human element.

As SMBs automate tasks and processes with AI, there is a danger of prioritizing efficiency and optimization over human considerations. A small coffee shop might implement an AI-powered ordering system to reduce wait times, but in doing so, they might inadvertently diminish the personal interaction between baristas and customers, which was a key part of their appeal.

Maintaining the human element in an AI-driven world requires a conscious effort to prioritize values such as empathy, compassion, and human connection. SMBs must consider how AI can augment, rather than replace, human capabilities. Perhaps AI can handle routine tasks, freeing up employees to focus on more meaningful interactions with customers.

Or perhaps AI can provide insights that enhance human decision-making, rather than automating it entirely. The challenge is to harness the power of AI in a way that strengthens, rather than undermines, the core human values that define SMBs.

Navigating the Moral Maze ● Practical Steps for SMBs

The moral implications of AI in SMBs Meaning ● AI empowers SMBs through smart tech for efficiency, growth, and better customer experiences. are not insurmountable obstacles. By taking proactive steps, SMBs can navigate this complex landscape responsibly and ethically. This begins with awareness.

SMB owners and employees need to educate themselves about the potential ethical risks associated with AI. This education should not be limited to technical aspects; it should encompass the broader societal and human implications of these technologies.

Next, SMBs should prioritize transparency and explainability when selecting and implementing AI systems. Choosing AI tools that offer insights into their decision-making processes, even if not fully transparent black boxes, is crucial. Engaging with vendors to understand the ethical considerations built into their products and demanding clear explanations of how AI systems function are essential steps.

Furthermore, establishing clear lines of accountability within the business for AI-related decisions and outcomes is paramount. This includes designating individuals or teams responsible for overseeing AI implementation Meaning ● AI Implementation: Strategic integration of intelligent systems to boost SMB efficiency, decision-making, and growth. and addressing any ethical concerns that arise.

Finally, SMBs must remain committed to their core human values. This means using AI in a way that augments human capabilities, rather than replacing them entirely. It means prioritizing fairness, equity, and compassion in AI-driven decisions.

And it means continuously evaluating the ethical impact of AI on customers, employees, and the community. The integration of AI into SMBs is not just a technological transition; it is a moral journey that requires careful navigation and a steadfast commitment to ethical principles.

Strategic Moral Integration Artificial Intelligence Small Businesses

The narrative surrounding artificial intelligence in small to medium-sized businesses often fixates on efficiency gains Meaning ● Efficiency Gains, within the context of Small and Medium-sized Businesses (SMBs), represent the quantifiable improvements in operational productivity and resource utilization realized through strategic initiatives such as automation and process optimization. and cost reduction. This focus, while understandable, obscures a deeper, more consequential reality ● the integration of AI is fundamentally reshaping the ethical terrain of SMB operations. We are not simply talking about automating tasks; we are confronting a paradigm shift in how SMBs make decisions, interact with stakeholders, and define their moral compass.

Beyond Compliance ● Ethical Strategy as Competitive Advantage

For many SMBs, ethical considerations are often relegated to compliance ● adhering to legal minimums and avoiding overt wrongdoing. AI compels a more proactive and strategic approach to business ethics. In an era where consumers are increasingly attuned to corporate social responsibility and ethical sourcing, SMBs that demonstrably prioritize ethical AI Meaning ● Ethical AI for SMBs means using AI responsibly to build trust, ensure fairness, and drive sustainable growth, not just for profit but for societal benefit. practices can cultivate a significant competitive advantage. This is not mere public relations; it is about building genuine trust and loyalty in a marketplace saturated with algorithmic uncertainty.

Consider two competing local bookstores. One implements AI-powered recommendation engines without considering potential biases in book selections or data privacy Meaning ● Data privacy for SMBs is the responsible handling of personal data to build trust and enable sustainable business growth. implications. The other bookstore, however, proactively audits its AI systems for bias, transparently communicates its data privacy policies to customers, and uses AI to promote diverse voices and authors. Which bookstore is more likely to attract and retain ethically conscious customers?

The answer is self-evident. Ethical AI is not a cost center; it is an investment in long-term sustainability and brand equity.

Algorithmic Auditing and Ethical Due Diligence

The black box nature of certain AI systems necessitates robust algorithmic auditing and ethical due diligence processes within SMBs. This is not solely the domain of large corporations with dedicated ethics teams. SMBs can and must develop practical strategies for assessing the ethical risks of their AI deployments. This begins with understanding the data used to train AI systems.

Where does the data come from? Does it accurately represent the diversity of the SMB’s customer base and community? Are there potential biases embedded within the data that could be amplified by the AI?

Furthermore, SMBs should implement mechanisms for monitoring AI system outputs and identifying potential ethical red flags. This could involve regular reviews of AI-driven decisions, feedback loops from employees and customers, and even external ethical audits. For instance, a small online retailer using AI for fraud detection should regularly review flagged transactions to ensure that legitimate customers are not being unfairly targeted.

Ethical due diligence is an ongoing process, not a one-time checklist item. It requires a commitment to continuous monitoring, evaluation, and refinement of AI systems.

Ethical AI is not just risk mitigation; it’s value creation for SMBs.

Table 1 ● Ethical Due Diligence Checklist for SMB AI Implementation

The Moral Tightrope ● Balancing Automation and Human Labor

AI-driven automation inevitably raises concerns about job displacement, particularly within SMBs where resources for retraining and redeployment may be limited. The moral challenge for SMBs is to navigate this automation imperative responsibly, balancing efficiency gains with the well-being of their employees and communities. This is not about resisting automation; it is about strategically managing its implementation in a way that minimizes negative social consequences and maximizes shared prosperity.

SMBs should proactively consider the potential impact of AI automation on their workforce. This might involve exploring opportunities for reskilling employees to take on new roles that complement AI systems, rather than simply replacing human labor. For example, a small manufacturing company automating its assembly line with AI-powered robots could invest in training its existing workforce to maintain and manage these robotic systems.

Furthermore, SMBs can explore alternative business models that leverage AI to enhance human capabilities and create new forms of value, rather than solely focusing on labor substitution. The future of work in SMBs is not predetermined; it is shaped by the ethical choices businesses make today.

Data Privacy and Security ● Building Trust in the Age of AI

AI systems thrive on data, and SMBs are increasingly collecting and processing vast amounts of customer and operational data. This data-driven paradigm intensifies the moral imperative for robust data privacy and security practices. Customers are entrusting SMBs with their personal information, and they expect this data to be handled responsibly and ethically.

Data breaches and privacy violations can erode customer trust, damage brand reputation, and expose SMBs to significant legal and financial risks. In the age of AI, data privacy is not just a legal requirement; it is a moral obligation.

SMBs must invest in cybersecurity measures to protect sensitive data from unauthorized access and cyberattacks. This includes implementing strong data encryption, access controls, and regular security audits. Furthermore, SMBs should be transparent with customers about their data collection and usage practices. Clear and concise privacy policies, readily accessible on company websites and in customer communications, are essential for building trust.

Obtaining informed consent from customers before collecting and using their data is a fundamental ethical principle. In the AI era, data stewardship is a core component of business morality.

Ethical AI Governance ● Embedding Values into Operations

To effectively address the moral implications of AI, SMBs need to establish clear ethical AI governance Meaning ● Ethical AI Governance for SMBs: Responsible AI use for sustainable growth and trust. frameworks. This is not about creating bureaucratic red tape; it is about embedding ethical considerations into the very fabric of business operations. Ethical AI governance Meaning ● AI Governance, within the SMB sphere, represents the strategic framework and operational processes implemented to manage the risks and maximize the business benefits of Artificial Intelligence. involves defining clear ethical principles that guide AI development and deployment, establishing processes for ethical review and oversight, and fostering a culture of ethical awareness throughout the organization. For SMBs, this can be achieved through practical and scalable measures.

This might include creating an internal AI ethics committee, even if it is composed of just a few employees from different departments. This committee can be responsible for developing ethical guidelines, reviewing AI projects for potential risks, and providing guidance to employees on ethical AI issues. Furthermore, SMBs can incorporate ethical considerations into their AI procurement processes, asking vendors about their ethical practices and demanding transparency about the ethical safeguards built into their AI products. Ethical AI governance is not a static document; it is a dynamic and evolving framework that adapts to the changing landscape of AI and business morality.

The Long View ● Sustainable and Responsible AI Growth

The integration of AI into SMBs is not a short-term trend; it is a long-term transformation. To ensure that this transformation is positive and beneficial for all stakeholders, SMBs must adopt a long-term perspective on ethical AI. This means moving beyond immediate efficiency gains and focusing on building sustainable and responsible AI Meaning ● Responsible AI for SMBs means ethically building and using AI to foster trust, drive growth, and ensure long-term sustainability. practices that align with broader societal values. Ethical AI is not just about avoiding harm; it is about contributing to a more just, equitable, and prosperous future.

SMBs can play a crucial role in shaping the ethical trajectory of AI. By prioritizing ethical considerations in their AI adoption strategies, they can demonstrate leadership and inspire larger organizations to follow suit. By sharing their experiences and best practices, they can contribute to a broader ecosystem of ethical AI innovation.

And by engaging in public discourse about the moral implications of AI, they can help shape policies and regulations that promote responsible AI development and deployment. The ethical future of AI is not solely determined by technologists and policymakers; it is co-created by businesses, including the vital SMB sector, and the choices they make every day.

Normative Frameworks Algorithmic Accountability Small Medium Enterprises

The deployment of artificial intelligence within small to medium-sized enterprises transcends mere operational optimization; it precipitates a profound re-evaluation of normative business ethics. The algorithmic agency inherent in AI systems necessitates a departure from conventional compliance-driven ethical postures toward a more robust, principle-based framework of algorithmic accountability. This transition is not simply about mitigating risks; it is about proactively constructing an ethical infrastructure that aligns AI implementation with the core values and societal responsibilities of SMBs.

Deontological and Utilitarian Perspectives on Algorithmic Morality

Traditional ethical theories offer divergent yet complementary lenses through which to analyze the moral implications of AI in SMBs. A deontological perspective, rooted in duty and moral obligation, emphasizes the inherent rightness or wrongness of actions, irrespective of their consequences. From this standpoint, SMBs have a categorical imperative to ensure that their AI systems adhere to fundamental ethical principles, such as fairness, transparency, and respect for human dignity. This necessitates rigorous pre-deployment ethical assessments and ongoing monitoring to guarantee algorithmic adherence to these non-negotiable moral duties.

Conversely, a utilitarian perspective, focused on maximizing overall well-being and minimizing harm, prioritizes the consequentialist outcomes of AI deployment. SMBs adopting a utilitarian approach must meticulously evaluate the potential benefits and harms of AI systems across all stakeholder groups ● customers, employees, suppliers, and the broader community. This requires a comprehensive cost-benefit analysis that extends beyond purely economic metrics to encompass social, ethical, and environmental impacts. For instance, while AI-driven automation Meaning ● AI-Driven Automation empowers SMBs to streamline operations and boost growth through intelligent technology integration. may yield short-term efficiency gains, a utilitarian analysis would also consider potential long-term societal costs, such as increased unemployment and economic inequality within local communities.

Algorithmic accountability in SMBs is not a technical challenge; it’s a philosophical imperative.

List 1 ● Contrasting Ethical Frameworks for SMB AI Implementation

• Deontology ●
  • Focuses on moral duties and principles.
  • Emphasizes inherent rightness or wrongness of AI actions.
  • Prioritizes fairness, transparency, and respect for human dignity in AI systems.
  • Requires pre-deployment ethical assessments and ongoing monitoring for principle adherence.
• Utilitarianism ●
  • Focuses on maximizing overall well-being and minimizing harm.
  • Prioritizes consequentialist outcomes of AI deployment.
  • Requires comprehensive cost-benefit analysis of AI impacts across stakeholders.
  • Considers social, ethical, and environmental consequences alongside economic metrics.

The Principalism Approach ● Autonomy, Beneficence, Non-Maleficence, Justice

Building upon these foundational ethical theories, the principalism framework, widely utilized in biomedical ethics, provides a pragmatic and actionable methodology for navigating the complex moral landscape of AI in SMBs. Principalism posits four core ethical principles ● autonomy, beneficence, non-maleficence, and justice. These principles offer a structured approach to ethical decision-making in AI, enabling SMBs to move beyond abstract ethical ideals toward concrete operational guidelines.

Autonomy, in the context of AI, underscores the importance of respecting individual agency and informed consent. SMBs deploying AI systems that interact with customers or employees must ensure transparency regarding data collection, algorithmic decision-making processes, and the potential impact on individual autonomy. This necessitates clear and accessible privacy policies, explainable AI Meaning ● XAI for SMBs: Making AI understandable and trustworthy for small business growth and ethical automation. interfaces, and mechanisms for individuals to exercise control over their data and algorithmic interactions.

Beneficence compels SMBs to utilize AI in ways that actively benefit stakeholders and contribute to the common good. This extends beyond mere profit maximization to encompass social value creation, community engagement, and positive societal impact. For example, an SMB could leverage AI to optimize resource allocation, reduce environmental footprint, or develop products and services that address unmet social needs.

Non-Maleficence, the principle of “do no harm,” mandates that SMBs proactively mitigate potential risks and harms associated with AI deployment. This includes rigorous bias detection and mitigation in AI algorithms, robust cybersecurity measures to prevent data breaches, and careful consideration of the potential for job displacement Meaning ● Strategic workforce recalibration in SMBs due to tech, markets, for growth & agility. and economic disruption caused by AI-driven automation.

Justice demands equitable distribution of the benefits and burdens of AI across all stakeholder groups. SMBs must ensure that AI systems do not perpetuate or exacerbate existing social inequalities, and that the advantages of AI are accessible to all, regardless of socioeconomic status, demographic background, or other differentiating factors. This requires a commitment to fairness in algorithmic design, inclusive AI implementation strategies, and proactive measures to address potential disparities in AI access and outcomes.

Algorithmic Bias Mitigation ● Statistical Parity Versus Equal Opportunity

A critical challenge in ethical AI implementation Meaning ● Ethical AI for SMBs: Strategic, responsible AI adoption for sustainable growth, balancing ethics with business needs. within SMBs lies in mitigating algorithmic bias. Two prominent approaches to fairness in machine learning, statistical parity and equal opportunity, offer distinct conceptualizations of algorithmic justice, each with its own strengths and limitations. Statistical parity, also known as demographic parity, aims to ensure that AI systems produce outcomes that are proportionally representative across different demographic groups. For example, in an AI-powered loan application system, statistical parity would require that the approval rate for loan applications is roughly equal across different racial or ethnic groups.

Equal opportunity, in contrast, focuses on ensuring that AI systems provide equal opportunities for positive outcomes to all qualified individuals, regardless of their demographic group. In the loan application example, equal opportunity would require that the AI system accurately identifies qualified applicants from all demographic groups, even if the overall approval rates differ due to underlying factors such as credit history or income levels. The choice between statistical parity and equal opportunity, or a hybrid approach, depends on the specific context, the nature of the decision being made by the AI system, and the ethical priorities of the SMB.

Table 2 ● Comparing Fairness Metrics in Algorithmic Bias Meaning ● Algorithmic bias in SMBs: unfair outcomes from automated systems due to flawed data or design. Mitigation

Explainable AI (XAI) and Algorithmic Transparency Imperative

The opacity of complex machine learning models, often referred to as the “black box” problem, poses a significant impediment to ethical AI governance in SMBs. Explainable AI (XAI) seeks to address this challenge by developing techniques and methodologies to make AI decision-making processes more transparent and understandable to human users. XAI is not merely a technical desideratum; it is a moral imperative for SMBs seeking to deploy AI ethically and responsibly. Algorithmic transparency Meaning ● Algorithmic Transparency for SMBs means understanding how automated systems make decisions to ensure fairness and build trust. is essential for accountability, trust-building, and effective human oversight of AI systems.

XAI techniques can range from simple rule-based explanations to more sophisticated methods that visualize decision pathways or identify key features influencing AI outputs. For SMBs, adopting XAI principles can involve selecting AI solutions that prioritize explainability, demanding transparency from AI vendors regarding algorithmic design and data usage, and implementing user interfaces that provide clear and understandable explanations of AI-driven recommendations or decisions. For instance, an SMB using AI for customer service chatbots could utilize XAI techniques to provide customers with insights into why the chatbot is recommending a particular solution or course of action.

Data Governance and Algorithmic Stewardship

Ethical AI implementation is inextricably linked to robust data governance Meaning ● Data Governance for SMBs strategically manages data to achieve business goals, foster innovation, and gain a competitive edge. and algorithmic stewardship practices. Data governance encompasses the policies, processes, and standards that govern the collection, storage, use, and sharing of data within an organization. Algorithmic stewardship extends these principles to the management and oversight of AI algorithms themselves. For SMBs, effective data governance and algorithmic stewardship are crucial for ensuring data privacy, security, quality, and ethical utilization in AI systems.

This involves establishing clear data access controls, implementing data anonymization and pseudonymization techniques to protect sensitive information, and conducting regular data audits to ensure compliance with privacy regulations and ethical guidelines. Furthermore, algorithmic stewardship requires ongoing monitoring of AI system performance, regular bias audits, and mechanisms for updating and retraining AI models to adapt to evolving data landscapes and ethical considerations. Data and algorithms are not neutral resources; they are ethically charged assets that require careful and responsible management.

Stakeholder Engagement and Participatory AI Design

Ethical AI governance in SMBs cannot be solely a top-down, technocratic endeavor. It requires meaningful stakeholder engagement Meaning ● Stakeholder engagement is the continuous process of building relationships with interested parties to co-create value and ensure SMB success. and participatory AI design processes that incorporate diverse perspectives and values. This includes actively soliciting input from employees, customers, community members, and other relevant stakeholders in the AI development and deployment lifecycle. Participatory AI design can help to identify potential ethical risks and unintended consequences early on, and to ensure that AI systems are aligned with the needs and values of the communities they serve.

SMBs can implement stakeholder engagement through various mechanisms, such as focus groups, surveys, public forums, and advisory boards. For example, a local retail store considering implementing AI-powered facial recognition for loss prevention could engage with community groups and privacy advocates to address concerns about surveillance and data privacy. Participatory AI design is not simply about ticking a box for stakeholder consultation; it is about fostering a collaborative and inclusive approach to AI innovation that prioritizes ethical considerations and shared societal benefit.

Long-Term Societal Impact and Existential Risk Mitigation

While the immediate focus of SMBs may be on the operational and ethical implications of AI within their own organizations, a broader ethical perspective necessitates consideration of the long-term societal impact Meaning ● Societal Impact for SMBs: The total effect a business has on society and the environment, encompassing ethical practices, community contributions, and sustainability. and potential existential risks associated with advanced AI technologies. This is not to suggest that SMBs are solely responsible for addressing these macro-level challenges, but rather that they have a role to play in contributing to a responsible and sustainable AI ecosystem. This includes supporting research and development in ethical AI, advocating for responsible AI policies and regulations, and fostering a culture of ethical awareness and critical reflection on AI technologies within their organizations and communities.

The potential for AI to exacerbate existing social inequalities, to displace human labor on a massive scale, and even to pose existential risks to humanity are not merely science fiction scenarios; they are legitimate concerns that warrant serious consideration. SMBs, as integral components of the global economy and social fabric, have a moral obligation to engage with these broader ethical challenges and to contribute to the development of AI in a way that aligns with human flourishing and long-term societal well-being. The ethical future of AI is not predetermined; it is shaped by the collective choices and actions of individuals, organizations, and societies around the world.

References

• O’Neill, Cathy. Weapons of Math Destruction ● How Big Data Increases Inequality and Threatens Democracy. Crown, 2016.
• Mittelstadt, Brent Daniel, et al. “The Ethics of Algorithms ● Current Landscape and Future Directions.” Big Data & Society, vol. 3, no. 2, 2016, pp. 1-21.
• Beauchamp, Tom L., and James F. Childress. Principles of Biomedical Ethics. 8th ed., Oxford University Press, 2019.
• Barocas, Solon, et al., editors. Fairness and Machine Learning ● Limitations and Opportunities. MIT Press, 2023.
• Doshi-Velez, Finale, and Been Kim. “Towards A Rigorous Science of Interpretable Machine Learning.” ArXiv Preprint ArXiv:1702.08608, 2017.