Artificial intelligence (AI) systems are evolving fast. However, ethical standards that ensure these systems don’t harm the public, such those that aim to prevent unintentional biases based on the data these systems are trained on, have been less quick to evolve. According to a global survey conducted by MIT Sloan Management Review, which polled over 1,000 executives, 82% of managers in organizations with at least USD $100 million in annual revenues agreed or strongly agreed that responsible AI (RAI) should be included in their top management agenda. At the same time, only 50% reported that RAI is a part of their top management’s agenda.
How can organizations that develop or use artificial intelligence ensure RAI is not just an afterthought? A recent panel of global AI experts, organized by MIT Sloan Management Review and global consulting firm BCG, concluded with the following takeaways:
- Leadership needs to understand why RAI is important to the organization’s strategy. Otherwise, RAI may never make it into the agendas of the organization’s major decision makers.
- Determine whether RAI is part of your AI strategy or a part of your wider organizational goals, such as corporate responsibility. Without an understanding of this, leadership may not fully grasp that it should be integrated into their larger agenda.
- Look at RAI as an urgent need that must be integrated now. Otherwise, you may miss valuable opportunities to prevent risk and harm down the line.
What are the Fundamental Principles of AI Ethics?
Understanding the core principles of AI is the first step to developing an effective AI standards framework. Such a framework should also align with an organization’s mission, as well as any regulations the organization may be affected by through its implementation of the AI system. According to TechTarget, the basic principles of ethical AI include:
- Fairness: The AI system does not contain biases and functions equally well for all groups
- Accountability: The AI system has ways to identify who is responsible across different stages of the AI life cycle if something goes wrong and provides ways for humans to supervise and control the system
- Transparency: When the AI system makes a decision, it allows humans to understand why it came to that conclusion, which is essential for building trust
- Safety: The AI system is equipped with effective security controls
What does incorporating these principles into an AI system look like in practice? During an interview with Analytics India Magazine, Layak Singh, CEO of Artivatic AI, an insurance platform, said the company reduces biases in AI by defining the business problems it wants to solve while considering end users, then configuring data collection methods to be able to incorporate diverse perspectives.
“We also ensure that we clearly understand our training data, as this is where most biases are introduced and can be avoided,” Singh said. “With that aim, we also ensure an ML [machine learning] team that’s assorted as they ask dissimilar queries and thus interact with the AI models in various ways. This leads to identifying errors before the model is underway in production and is the best manner to reduce bias both at the beginning and while retraining models.”
Additionally, there is a major focus on feedback as his company keeps feedback channels, such as forum discussions, open in order to run continual audits and upgrades.
Ensuring AI systems are ethical is becoming essential to building trust with clients and customers. Don’t wait until that trust is already broken— start developing an ethical AI standards framework today.
Incorporating AI Standards at Your Organization
An online five-course program, AI Standards: Roadmap for Ethical and Responsible Digital Environments, provides instructions for a comprehensive approach to creating ethical and responsible digital ecosystems. Contact an IEEE Content Specialist to learn more about how this program can benefit your organization.
Interested in getting access for yourself? Visit the IEEE Learning Network (ILN) today!
Krishna, Sri. (20 April 2022). Talking Ethical AI with Artivatic’s Layak Singh. India Analytics Magazine.
Kiron, David, Renieris, Elizabeth, and Mills, Steven. (19 April 2022). Why Top Management Should Focus on Responsible AI. MIT Sloan Management Review.
Kompella, Kashyap. (1 April 2022). How AI ethics is the cornerstone of governance. TechTarget.