Firms turn cautious over gen AI amidst fears of obsolescence, fragmented regulations

Of 47 APAC jurisdictions, about 16 have AI regulations or guidance in place.

Companies in the Asia Pacific region, including banks, are proceeding cautiously over their generative AI (gen AI) plans in fear that their products could become outdated fast.

“Our clients set aside funds for Generative AI but proceed cautiously due to its evolving nature and current limitations. Firms are concerned their products could quickly become outdated amid ongoing AI advancements,” Chris Leung, EY Asia Pacific FSO and partner, technology consulting, told Asian Banking & Finance via written correspondence.

Products could reportedly become outdated in as fast as a few months, especially with the rapid development of AI agent concepts and models.

But banks also cannot disengage fully: AI-enabled digital workers are transforming operations, handling low-value tasks like simple email replies and calendar management to larger tasks like mitigating financial risks and even ensuring regulatory compliance.

This way, human employees can instead focus on making more complex decisions.

Even in decision-making, AI is expected to play a role, said Siobhan Byron, EVP Universal Banking, Finastra Limited.

“Another emerging trend is AI-driven analytics and intelligent dashboards. These provide banks with rapid insights into customer needs, successful strategies, and areas for growth, leading to personalized customer experiences,” Byron told Asian Banking & Finance via written answers, in response to a question about emerging trends in the banking industry for 2025.

Another application is GenAI-powered assistants, which Byron expects to be implemented both within and outside of banks. This should support informed decision-making for banks and their customers, she said.

Regulatory challenges
Regulatory compliance is one factor that appears both as a possible use case and a challenge for companies in the APAC: with not less than 47 countries and territories, companies must grapple with varying data privacy and residency laws.

As of 2024, 16 of the not less than 47 jurisdictions in APAC have some form of AI guidance or regulations, according to American multinational legal firm Sidley Austin LLP.

Some countries are implementing AI-specific laws and regulation, while others take a “softer” law approach in reliance on nonbinding principles and standards, the law firm said in an August 2024 report.

India, for example, has launched an AI advisory group. Japan and many Southeast Asian countries including Indonesia, Malaysia, Thailand, and Vietnam, instead opted for AI laws. In contrast, Singapore had no laws at the moment but instead launched a generative AI framework in May 2024.

China and Japan, meanwhile, have stringent regulations that often prohibit cross-border data exchange.

This fragmentation equates to higher costs and complexity for banks, according to Leung.

“The acceleration in adoption of AI is increasing the risk of non-compliance with data privacy regulations and the likelihood that an organization could inadvertently damage customer confidence,” he said.

Banks are further advised to develop their own AI governance frameworks, to enable compliance with applicable legal and regulatory requirements, Sidley Austin LLP said in its report.

The good news is that, as AI regulations become more complex in APAC, they might actually become similar if not uniform with one another, according to a separate report by Sia Partners published in July 2024.

“There is an anticipation of uniformity predicated on historical instances of countries aligning with EU directives,” the Paris-headquartered management consulting firm said.

For example, Australia’s government has indicated they may adopt a risk-based approach like the EU AI Ac.

“This move may facilitate smoother international business operations and standardise ethical AI usage across borders, a trend that is expected to extend to the APAC region,” it added.

For now, future legislation in APAC is expected to lean towards enforcing binding regulations and penalties, without further legislative action, said Sidley Austin LLP.

Separately, the APAC region boasts a vibrant start-up and FinTech community, offering numerous innovative use cases for AI. This dynamic environment fosters creativity and rapid development in AI technologies, according to Leung.

In particular, Asia's principles-based regulatory approach to AI fosters a competitive market, enabling firms to experiment with GenAI technology within flexible guidelines, driving innovation in financial services, he said.

“The supportive regulatory environment in Asia, guided by principles like fairness, ethics, and transparency, positions the region as a leader in GenAI implementation, leveraging existing frameworks to address new risks and drive responsible AI development,” Leung said.

Effective AI regulation hinges on specific guidance and sector-tailored rules, highlighting the importance of harmonized standards and cross-border collaboration to prevent regulatory fragmentation and build global trust in AI systems.

“Embracing human-in-the-loop practices, supporting regulatory sandboxes, and prioritizing consumer protection and ethics are key steps for regulators and organizations to navigate the evolving AI landscape, fostering innovation while upholding safety and reliability,” Leung said.

— with addition reporting by Jaleen Ramos