Data Leaders members recently came together to discuss the governance hurdles and the urgent need to source solutions to address the associated organisational, technical and regulatory concerns.

Data Leaders members can read the full post and connect with peers via the Data Leaders Hub.

Organisational Challenges

The Rush to Innovate

There’s an undeniable enthusiasm for GenAI across sectors, leading to a surge in proposed use cases. This pressure to deliver quickly poses governance challenges, encouraging hasty decision making without thorough consideration of data privacy, ethics, or security.

CDOs urge data professionals to focus on being good rather than being first to mitigate errors that could damage the company’s organisation or have more serious repercussions later down the line.

Decentralised Experimentation

As departments experiment with GenAI independently, the lack of centralised oversight can result in inefficiencies and governance lapses. Each project may operate under its own set of rules, overlooking essential governance aspects.

Central coordination is vital to ensure a unified approach that safeguards against disjointed and risky practices.

Prioritising Use Cases

Effective governance is also hampered by the challenge of use case prioritisation. Without a strategic approach to selecting projects based on factors like budget, data sensitivity, and potential business impact, resources may be squandered on misaligned initiatives.

A focused strategy ensures that GenAI applications contribute positively to organisational objectives while being manageable from technical and regulatory standpoints.

Technical Challenges

Navigating Technical Complexities

GenAI projects often vary in technical complexity. Some use cases may require the development of advanced machine learning models, complex algorithms, and sophisticated data processing techniques. Additionally, integrating GenAI solutions with existing systems and workflows can introduce technical intricacies.

Here, clear governance structures including roles and responsibilities, guidelines and standards for model development are essential to consistently validate, deploy and manage them.

Data Quality and Accuracy

Ensuring the quality and availability of data for GenAI projects is both a fundamental technical and organisational challenge. Poor data quality, inadequate data availability, or data bias pose significant threats to developers, users, decision-makers, and ultimately the organisation.

It requires dedication to seeking out the best and emerging solutions to reliably and quickly check and improve data quality, while embedding an organisational governance structure to infuse a data-first mindset.

Regulatory Uncertainties

Unclear Roles and Responsibilities

While each company’s organisation is unique, it’s evident that there is ambiguity in most organisations regarding who should be responsible for different aspects of GenAI project; should IT teams take the lead, or should business units drive the process? This lack of clarity has serious implications for governance, creating gaps when it comes to data management, privacy and ethics.

To address this challenge, data professionals must define and communicate clear roles and responsibilities within each GenAI initiative to ensure that every aspect of governance is covered and establish centralised AI data governance roles, frameworks and policies as soon as possible.

Bridging the Knowledge Gap

Another widely recognised challenge of effective governance is the lack of data knowledge among decision makers and gatekeepers, particularly legal whose prime concerns often relate to compliance of data privacy regulations such as GDPR and other regional data protection laws. While some CDOs have successfully forged effective collaborations with legal, others have not yet been able to due to a lack of understanding of the technical aspects of GenAI. This can hinder decision making, as they struggle to assess the risks and benefits accurately.

It’s clear from the discussion that data professionals view the lack of governance as the key challenge and concern to safely deploying GenAI at scale. It also seems that smaller organisations and those with more mature data governance organisations are more likely to successfully collaborate with principal stakeholders to mitigate risks while those in less mature organisations may feel pressured to move before they’re ready.

Ensure you get stories like this and many more interesting insights from data and analytics leaders like you – directly to your inbox, twice a month – by signing up to our newsletter. Would like to become a member? Get in touch!