Malaysia's Social Security Organisation (Perkeso) has successfully identified fraudulent claims in its Daya Kerjaya 2.0 employment incentive programme through a dual-track approach combining artificial intelligence technology with information supplied by internal and external whistleblowers. The organisation's leadership revealed this enforcement strategy as part of broader efforts to safeguard public funds and ensure that financial support reaches genuinely eligible participants.

The deployment of AI-driven analysis represents a modernisation of Perkeso's fraud detection capabilities, moving beyond traditional manual review processes that are often labour-intensive and prone to human error. By utilising machine learning algorithms capable of analysing vast datasets at scale, the organisation can identify suspicious patterns, inconsistencies, and anomalies that might indicate fraudulent activity. This technological approach allows investigators to prioritise cases warranting deeper scrutiny, thereby allocating limited investigative resources more efficiently toward cases with higher risk profiles.

Whistleblower reports have emerged as a critical complement to automated systems, providing on-the-ground intelligence that algorithms alone cannot generate. Individuals with direct knowledge of fraudulent practices—whether employees within participating companies, fellow programme beneficiaries, or concerned members of the public—have furnished information that has triggered formal investigations. This crowdsourced vigilance recognises that fraud detection often depends on human awareness of circumstances that patterns in administrative data might not easily reveal. The combination acknowledges that technological solutions and human insight each carry distinct advantages, and their integration produces more robust outcomes.

The Daya Kerjaya 2.0 programme, which provides employment incentives to encourage business hiring and workforce development, carries substantial public investment and political significance. As a key component of Malaysia's employment support framework, the scheme's integrity directly affects public confidence in government initiatives designed to stimulate job creation. Fraudulent claims that divert funds to ineligible parties not only waste public resources but also undermine the programme's intended impact on legitimate job-seekers and employers participating in good faith.

Fraud in employment incentive schemes often takes several forms: companies inflating the number of workers hired or falsifying employment records; individuals claiming benefits while already employed elsewhere; misrepresentation of worker qualifications or circumstances; and organised schemes involving collusion between multiple parties. Each variant requires different investigative approaches, and the combination of AI and whistleblower intelligence allows Perkeso to detect multiple fraud typologies rather than relying on detection methods effective only against specific schemes.

The reliance on whistleblowers also underscores the importance of establishing clear, accessible reporting channels and providing protection to those who come forward. Organisations that successfully combat internal fraud typically implement robust protections against retaliation, clearly communicated reporting procedures, and demonstrated commitment to acting on credible information. Perkeso's apparent success in generating whistleblower reports suggests the organisation has established sufficient trust and institutional mechanisms to encourage disclosure despite potential professional or social risks that tipsters might face.

From a Southeast Asian perspective, Malaysia's approach reflects broader regional trends toward leveraging technology in government service delivery and fraud prevention. Neighbouring countries grappling with similar challenges in employment support schemes and social security administration increasingly recognise that traditional bureaucratic controls prove insufficient in environments where fraudulent schemes grow more sophisticated. The experience documented here may offer practical lessons for other ASEAN nations designing or upgrading their own social protection systems.

The AI implementation also raises important questions about data governance and privacy that Malaysian policymakers continue navigating. While powerful algorithms enable fraud detection, their application must balance security concerns against individuals' rights to privacy and protection against false accusations. The role of AI systems in generating investigative leads—as opposed to making final determinations—helps maintain human oversight and accountability in enforcement decisions, though ongoing evaluation of algorithmic fairness remains crucial.

Looking forward, Perkeso's experience suggests that combating fraud in employment incentive programmes requires sustained, multi-layered approaches rather than one-time interventions. As fraudsters adapt their tactics to evade detection, enforcement strategies must similarly evolve. Continuing investment in AI capabilities, maintaining strong whistleblower mechanisms, and ensuring coordination between different investigative functions will likely determine Perkeso's long-term success in protecting programme integrity. The organisation's willingness to acknowledge and publicise its anti-fraud efforts also serves a deterrent function, signalling that fraudulent claims carry genuine risk of detection and consequences.

For Malaysian employers and workers legitimately participating in Daya Kerjaya 2.0, these enforcement measures ultimately serve their interests by preserving programme resources and maintaining its credibility with policymakers. Programmes undermined by uncontrolled fraud face pressure to restrict eligibility, reduce benefit levels, or terminate entirely, outcomes that harm genuine beneficiaries. By implementing sophisticated fraud detection and enforcement, Perkeso supports the programme's long-term sustainability and effectiveness as an employment support tool for Malaysia's labour market.