Trending job role in AI: Governance & Policy Lead

The year started with some interesting recruitment news in the area of AI responsibility. Please take a note, this is one example of the new job roles created by the rise of AI. I’ve been following very actively the emerge of this new role, and thus wanted to post a short overview on the role of “AI governance and policy lead”, and explain how we at Saidot support these people do great in their important roles.

So let’s take a look how NYC Mayor’s Office outlines the responsibilities of an ALGORITHMS MANAGEMENT AND POLICY OFFICER (AMPO):

“Establish governing principles to guide City agencies in balancing the ethical and innovative uses of algorithmic systems and tools in agency decision-making, to ensure they provide the greatest benefit for New Yorkers and the City;”

Ethics principles is the natural first step for every AI governance lead. Many organizations have it already, but if not, AI governance lead should not waste time getting there. Principles outline the values that are important to the organization, and they should be committed to by the top management. I’m very happy to see both the ethics and the innovation being highlighted here.

With Saidot we build on our experience from helping tens of organizations in private and public sectors in setting principles for AI ethics and governance. We maintain best practice governance framework and processes to help AI governance leads take a jump start in creating their own frameworks.

“Design and implement a framework to help agencies identify, prioritize, and assess algorithmic tools and systems that support agency decision-making, considering their complexity, the benefits, impact, and any potential risk of harm.”

With an AI governance framework, we formalize the way in which we manage our algorithmic systems. This sets clear and systematic way for turning principles into operational requirements, processes and responsibilities. While companies deploy more and more algorithmic systems into production, many have already faced the need for a proper governance frameworks for managing the quality, risks and compliance of them.

“Develop and implement policies and protocols to guide the City and its agencies in the fair and responsible use of such tools and systems, considering the unique mission, purpose, and operational needs of each agency;”

Policies and protocols standardize the different aspects of the governance framework, such as transparency, risk & impact assessments, in a systematic way across agencies or departments, and even between various organizations. One can’t overestimate the importance of cross organizational policies while ensuring efficiency of governance in an ecosystem context.

“Design and implement protocols for agency reporting to the AMPO on the use of algorithmic systems and tools;”

Every AI governance & policy lead must establish a mechanism for getting up-to-date reporting on their portfolio of algorithmic systems. In practice, this is about managing the inventory or register of algorithmic systems, and having business friendly reporting across the whole portfolio.

All of the above requirements have been essential in shaping Saidot technology focus. We develop technology platform to provide our customers with

• Register for algorithmic systems. Metadata models for industry specific transparency, risk and impact assessment policy and compliance.
• Online reports for business stakeholder and human operator reporting.
• Models and standardized reports for end-user explainability and consumer & community informing.
• Audits and certifications for independent validation & oversight.

“Plan and implement a public engagement and education strategy related to the City’s use of algorithmic tools and systems; Create and maintain a public-facing platform that provides a mechanism for interacting with the public on algorithmic systems and tools and providing resources, including fielding questions, complaints, and concerns; advising agencies on addressing these.”

Any AI system to operate successfully must gain the trust of the people influenced by it. Openness and inclusion are critical elements for maintaining trust between city and its local community.

Organizations must prepare for explaining the logic of their algorithmic reasoning for the influenced end users. These explanations must cover both the model level explainability, as well as explanations on why system came to a specific decision or outcome (local explainability). For a public organization like city, this is vital requirement in align with acts of openness of government activities.

“Research new developments and best practices in managing the City’s use of algorithmic tools and systems in agency decision-making, and remain current in this emerging field;”

AI governance is a quickly developing domain and AI policy leaders must find ways for creating clarity while keeping themselves up to date on new developments on regulation, governance best practices and technical tools. In Saidot we contribute regularly to research and policy making on AI governance, and connect our customers with leading experts and data scientists focused on AI explainability.

I hope this gave you some interesting and practical insight into everyday life of an AI governance lead. If you got interested, New York City Mayor’s Office takes in applications on their AMPO role here . If you already have an interesting role in AI governance, let’s catch up for working it out! Some of you may already work in these roles, so, please jump in and comment.

Meeri