Common Risks of Machine Learning | Singapore Government Developer Portal
Have feedback? Please

Common Risks of Machine Learning

28 Oct 2021

In order to cater to the demand for complex interconnected systems, Machine Learning (ML) systems have evolved from traditional IT systems, which are typically simple programmes that handle simple datasets. ML systems create more opportunity by enabling organisations to process more complex use cases which are typically done by humans.

FEATURES OF ML SYSTEMS

ML_intro ML_intro

MITIGATING THE RISKS OF USING ML SYSTEMS

The table below describes key risk areas of using ML systems and possible mitigation measures to reduce associated residual risks.

ML_intro

KEY CYBERSECURITY FACTORS AND APPROACHES

ML systems process and distil massive amounts of data, reinforcing the need for fundamental cybersecurity measures to protect data confidentiality and integrity, as well as to maintain the availability of ML services. Organisations may consider the following key factors and cybersecurity approaches:

ML_intro