In the ever-evolving landscape of artificial intelligence (AI), ensuring fairness, transparency, accountability, and robustness has become paramount. To achieve these goals, organizations are increasingly establishing in-house centers of excellence dedicated to trustworthy AI practices. These centers serve as hubs of knowledge and collaboration, where individuals from diverse backgrounds come together to champion ethical AI. In this article, we will outline ten crucial steps for creating an effective and impactful AI Center of Excellence within your organization.

What’s the Problem?

AI bias is a general concept that refers to the fact that an AI system has been designed, intentionally or not, in a way that may make the system’s decisions or use unfair. Bias can be present both in the algorithm of the AI system and in the data used to train and test it. AI bias can emerge in an AI system as a result of cultural, social, or institutional expectations because of any of the following reasons:

1. Design limitations that are technical in nature
2. Use in unanticipated contexts 
3. By making decisions about stakeholders that are not part of the initial design 

Fairness within an AI system is general concept that refers to the equitable treatment of individuals or groups of individuals. The choice of a specific notion of fairness for an AI system depends on the context in which it is used. One possible reason for an AI system to behave unfairly is the presence of bias in the training data.

To practice trustworthy or responsible AI, a number of organizations are creating in-house centers of excellence. These are groups of trustworthy AI stewards from across the business that can understand, anticipate, and mitigate any potential problems. The intent is not to necessarily create subject matter experts but rather a pool of ambassadors who act as focal points.

01 Strategically Forge Groundbreaking Bonds 

Begin by identifying pockets of interest and expertise related to AI and AI ethics within your organization. Bring together these disparate groups, whether they are from different geographical locations or various disciplines, into a unified space for sharing insights and knowledge. Consider leveraging digital platforms like Slack or online communities to facilitate discussions and information exchange. For example, engage minority groups with a vested interest in AI ethics to collaborate with data scientists addressing bias concerns.

02 Flatten Hierarchy 

A successful AI Center of Excellence thrives on collaboration and inclusivity. Embrace a flat hierarchy where everyone’s voice is heard, and leadership roles rotate. This coalition of changemakers should foster an environment where all members support one another. The approach must engage stakeholders from across an organization, from data scientists and Chief Technology Officers (CTO) to Chief Diversity and Inclusivity Officers. Fighting bias and ensuring fairness is a challenge that is solved by more than just good technology and by more than just one kind of stakeholder.

03 Harvest Your Power 

Identify potential AI ambassadors from within the Center of Excellence. These ambassadors will play a crucial role in operationalizing trustworthy AI principles throughout the organization. Their responsibilities may include explaining the AI lifecycle, infusing ethics into design thinking, creating feedback loops, and advocating for adversarial testing. Ensure that these ambassadors are excellent communicators who can articulate the importance of ethical AI practices.

04 Begin Teaching Trustworthy AI Training at Scale 

Prioritize the education of your entire workforce on trustworthy AI practices. Develop tailored learning modules that cater to the specific needs and roles of various competency centers within your organization. 

05 Collaborate with Diverse Stakeholders 

Encourage your AI ambassadors to bridge silos within your organization. Invite stakeholders from diverse backgrounds, including diversity and inclusivity, HR, data science, and legal counsel, to collaborate. This interdisciplinary approach ensures a comprehensive understanding of AI ethics and promotes responsible AI across the organization.

06 Establish Clear Governance 

Define and publish governance standards for trustworthy AI that your organization adheres to. This serves as a foundational document outlining the principles and guidelines for ethical AI practices. Transparency in governance builds trust both internally and externally.

07 Promote Diversity and Inclusion 

Diversity is a cornerstone of trustworthy AI. Ensure that your AI teams are as diverse and inclusive as possible. Diverse perspectives help identify biases and promote fairness in AI models. Create an environment where all voices are valued and heard.

08 Implement Bias Mitigation Tools 

Introduce tools and AI engineering practices that aid in bias detection and mitigation. These tools are essential for ensuring that AI systems are fair, accountable, and robust. Make it a part of your AI development process to actively address bias and discrimination. These requirements can be interpreted as a need for platforms that can integrate with external systems to ingest large data sets, curate them, and predict based on deployed machine learning (ML) models.

09 Collaborate with an external AI Ethics Board

Identify, build, and collaborate with an external AI ethics board or advisory groups. These experts can provide valuable guidance on complex ethical dilemmas and ensure that your AI initiatives align with ethical and international standards. Some useful examples are shown below:

The Executive Committee for IEEE

Global Initiative on Ethics of Autonomous and Intelligent Systems

Linux Foundation AI

The International Organization for Standardization (ISO)

The World Economic Forum

CompTIA AI Council

MIT-IBM Watson AI Lab

The International Committee for Information Technology Standards (INCITS)

10 Champion Ethical Storytelling 

Encourage your AI ambassadors to be compelling storytellers. They should convey the importance of ethical AI practices to various stakeholders, from employees to customers. Storytelling can create a shared narrative that reinforces the significance of responsible AI within the organization.

Conclusion

Establishing a fair and trustworthy AI Center of Excellence is a crucial step for any organization venturing into AI. By following these ten steps, you can create a culture that embraces trustworthy AI practices that will become ingrained in the organization’s DNA, ensuring that AI benefits all and avoids unintended consequences.