Table of Contents
1. Please tell us your name, a little more about yourself and your experience so far in the data and AI industries.
Mrs. Sukanya Konatam is an IT professional boasting an extensive career of over 18 years, enriched by five years dedicated to specializing in AI and AI Governance. Her expertise spans a comprehensive range of IT disciplines, underscored by a profound depth of knowledge in the governance of artificial intelligence. This unique combination of experience positions her as a leading figure in the field, adept at navigating the intricacies of AI technology with a strategic and ethical approach. She has implemented data-centric solutions for several industries including banking & financials, telecom, health care, automobile, criminal justice and many more.
Currently, Sukanya is leading Enterprise Data Governance Modernization efforts at Vialto Partners with multiple teams of data engineers, data scientists & analysts. With her proven record of strategic thinking and Solutions Architecture Design, she is particularly responsible for creating a 360-degree view for the system, data security, mastering data management, enriching data sets, building modern dashboards and automating manual data processes. She defines indicators of performance and quality metrics, ensures compliance with data, determines roles and responsibilities related to data governance and ensures clear accountability. She facilitates the development and implementation of data quality standards, data protection standards and adoption requirements across Vialto Partners.
Prior to her role at Vialto Partners, Sukanya was leading data science, data analytics, automation, and administration teams at EFG Corporate Finance that supported each line of business, CFO and the operating committee with Financial Reporting, Budget & Forecasting automation, and Strategic Oversight. Sukanya helped create periodic KPI Metrics reports for the management and these were invaluable in helping identify new business opportunities pertaining to the use of information assets to achieve efficiency and effectiveness. By representing data as a strategic business asset at the senior management table, these reports proved key for critical decision-making to gauge business-critical initiatives, objectives, or goals. She was responsible for rolling out an enterprise-wide data governance framework, with a focus on the improvement of data quality and the protection of sensitive data through modifications to organization behavior policies and standards, principles, governance metrics, processes, related tools, and data architecture.
2. What are the challenging aspects in AI Governance?
In the realm of AI governance, certain areas remain enigmatic and pose challenges. These zones are marked by ambiguity, complexity, and ethical dilemmas, which can complicate decision-making when it comes to AI systems. Those are:
1. Transparency Standards:
- Determining to what extent an AI system should explain its decisions to users or stakeholders.
- Balancing transparency with the need to protect proprietary information and address security concerns, simultaneously.
2. Equity Assessment:
- Evaluating and mitigating biases within AI algorithms.
- Defining the concept of fairness and implementing it effectively.
3. Safety Precautions:
- Ensuring the safety and reliability of AI systems.
- Addressing unforeseen risks and unintended consequences.
4. Human-AI Collaboration:
- Clearly defining the roles and responsibilities of humans and AI in various contexts.
- Striking a harmonious balance between automation and human judgment.
5. Liability Frameworks:
- Determining who is responsible when AI systems cause harm or make errors.
- Establishing legal and ethical accountability.
To ensure responsible and beneficial AI deployment, it is crucial to collaborate with governments, civil society, and AI practitioners in crafting context-specific guidelines and policies.
3. What are the solutions to address the challenging aspects in AI governance?
To navigate the ambiguous aspects in AI governance, we need a comprehensive approach that involves collaboration, research, and policy formulation. Here are some strategies to address these challenges:
1. Transparency Standards:
- Develop clear guidelines for AI systems to explain your decision-making processes.
- Encourage research on interpretable AI models and techniques.
- Foster transparency by sharing information about data sources, model architectures, and training processes.
2. Equity Assessment:
- Establish standards for assessing and mitigating biases in AI algorithms.
- Regularly audit AI systems for fairness and address any disparities.
- Promote diverse teams in AI development to reduce bias.
3. Safety Precautions:
- Create safety frameworks for AI deployment.
- Conduct risk assessments to identify potential harms and unintended consequences.
- Encourage responsible testing and validation of AI systems.
4. Human-AI Collaboration:
- Define clear roles for humans and AI in decision-making.
- Design AI systems that complement human expertise rather than replace it.
- Foster interdisciplinary collaboration among AI researchers, ethicists, and domain experts.
5. Liability Frameworks:
- Establish legal frameworks to assign responsibility for AI outcomes.
- Consider liability insurance or risk-sharing mechanisms.
- Encourage ethical guidelines for AI practitioners.
- Involve the public in discussions about AI governance.
- Educate policymakers, businesses, and the public about AI risks and benefits.
- Foster a culture of responsible AI use.
- Collaborate across borders to create consistent AI governance standards.
- Share best practices and learn from different regulatory approaches.
Keep in mind that navigating these ambiguous areas is an ongoing journey, necessitating a delicate equilibrium between innovation and conscientious implementation. Through collaborative efforts, we can shape AI governance to enhance society while mitigating risks.
Conclusion:
In the realm of AI governance, ambiguous aspects lead to ongoing debates regarding where the boundaries should be drawn in terms of permissibility and responsibility. Reasonable arguments exist on both sides. For instance, how should society weigh the advantages of AI-powered surveillance for crime prevention or locate missing individuals against the implications it poses for privacy and human rights?
Please find my other publications in Data Governance here:
Data Governance: MDM and RDM (Part 3) – DZone
Data Governance: Data Architecture (Part 2) – DZone
Data Governance, Data Privacy and Security – Pt.1 – DZone
Please connect with me on LinkedIn Sukanya Konatam | LinkedIn