2023 was the year that Generative AI took center stage, with many individuals and organizations trying their hand at it and using it in various business settings. The time and labour that Generative AI saves are vast - in an internal study, we found that Generative AI boosted productivity by 62 percent, translating to 26 hours saved per week.
Many corporations are becoming aware of these benefits and are looking to integrate Generative AI within organizational workflows, and into various facets of business operations. However, with this comes the need for governance training.
Uncovering the risks in Generative AI that are raising concerns for better governance
AI has permeated every industry, from healthcare and finance to manufacturing and beyond. The recent expansive use of the latest Generative AI brings forth ethical dilemmas. Bias in algorithms, privacy concerns, and unintended consequences have begun to surface and these are challenges that data professionals must now grapple with. Therefore, it is necessary for Data Governance professionals to be educated so that they can guide implementations and deployments responsibly and ethically.
Global convergence of AI regulations and frameworks signals urgency
On a global scale, there is active action by various regulators and conversations on the governance of AI and its impact on people, jobs and privacy. This focus on AI regulation points to the urgency and also a convergence or alignment of frameworks for effective international coordinated standards.
Across ASEAN, whilst AI has been identified as a key focus area in national strategy, there are largely no AI governance regulations in place yet. In Malaysia, a set of AI governance and code of ethics that would form the basis of AI regulation, will be introduced this year. It is expected to include the Seven Principles of Responsible AI as outlined in the country’s National Artificial Intelligence Roadmap 2021-2025, as well as emphasise education and ethics. The Philippines plans to base a proposal for an ASEAN-wide AI regulatory framework, based on its draft legislation on AI, when it chairs the bloc in 2026.
Meanwhile, Singapore has been at the forefront of embracing AI technologies but acknowledges the need for responsible governance. In 2019, the first National AI Strategy was rolled out, outlining plans to deepen the use of AI to transform the economy. The latest National AI Strategy 2.0, unveiled in December last year, places a strong emphasis on industry collaboration, research and development, talent development, and establishing an effective and trusted environment for AI innovation.
Furthermore, the country actively participates in global conversations on AI. For example, it is a founding member of the Global Partnership on Artificial Intelligence. It has also been working with the United States to find areas of alignment between the US AI Risk Management framework and Singapore’s Governance framework.
The risks of Generative AI and personal data protection considerations
Generative AI brings value, but it is also important to be aware of its risks and the constraints. For instance, the transition from content creation to content generation is expected to increase privacy and security risks, as well as ethical-related breaches. These can be brought on by malice, accident, or ignorance in the use of Generative AI.
We are already seeing deep fakes and voice cloning, and these understandably raise concerns about the rampant spread of misinformation, data privacy and intellectual property. This is likely to increase in 2024, which could be very damaging, particularly in the political arena - there are 80 major elections anticipated globally. Recently in Indonesia, an AI-generated avatar of the late political leader Suharto was created and disseminated in an attempt to influence voter choice at the upcoming Indonesian elections.
Rise of custom GPTs opens the door for information leaks and copyright issues
OpenAI opened its GPT store in January 2024. This has enabled users to create custom GPTs and upload files as part of its knowledge. There are currently no guardrails on uploading copyrighted content, which can be troubling for unknown or less governed developers who could monetize these GPTs in future. The ease of generating such custom GPTs can also bring the risk of adversarial prompts.
These include prompt injection (inserting malicious content to manipulate the output), prompt leakage (unintentional disclosure of sensitive information in responses), and jailbreaking (tweaking prompts to bypass AI system restrictions). These manipulations can have far-reaching consequences, and thus it is paramount that these challenges are addressed.
It still is very necessary to have the ‘human in the loop’ to supervise, fact check and validate the technology’s output. Issues such as potential biases and security vulnerabilities like data leakage of personal data, must be kept in mind by data professionals when they create a framework for the deployment of Generative AI.
Data professionals must upskill
Singapore’s Deputy Prime Minister Lawrence Wong said, when announcing an updated National AI Strategy in December 2023, that Singapore intends to triple its AI talent pool in three to five years to 15,000 by training locals and hiring overseas.
As Generative AI moves into mainstream, the role of data professionals will evolve from mere custodians of information to guardians of ethical AI practices. Thus, there is a crucial need for data professionals to prioritize AI governance training, instead of taking a ‘wait-and-see’ stance. As AI becomes more integrated into decision-making processes, data professionals must possess the skills to navigate the ethical complexities surrounding data management. As such, Data Protection Officers (DPOs) and Data Governance Professionals will be invaluable as firms develop and adopt Generative AI.
