From safety to security: preparing for an autonomous future

As robots enter public service, securing their software and supply chains becomes essential to keeping people safe. ST Engineering’s Goh Eng Choon explains why that requires independent testing before deployment and continued monitoring throughout their working lives.

Autonomous robots and other connected systems are becoming more common in public and operational settings, raising new questions around how they are secured throughout their lifecycle. Image: Canva

A robot delivering parcels or patrolling a public pathway needs to do more than avoid collisions. It also needs to resist attempts to hijack its controls.

 

That challenge is becoming tangible in Singapore's Punggol Digital District, where a physical AI testbed will see autonomous robots operating alongside residents in shared public spaces, performing everyday urban services.

 

As machines begin working alongside the people they serve, cybersecurity becomes a matter of physical safety.

 

The global market is still small but growing rapidly. About 7,000 humanoid robots were sold worldwide in 2025, according to the International Federation of Robotics.

 

Counterpoint Research tracked more than 22,000 shipments in the first half of 2026 alone, while Bank of America projects annual shipments could reach 10 million units by 2035.

When cyber threats become physical risks

 

Humanoid robots are among the newest generation of autonomous cyber-physical systems.

 

They rely on cloud services, fleet management platforms, software updates, and connected devices.

 

These connections enable powerful capabilities but are also potential attack surfaces that can affect how a platform perceives, decides, and acts.

 

"The challenge is no longer securing a device. It is securing an entire cyber-physical ecosystem," says Goh Eng Choon, President, Cyber, ST Engineering.

 

Goh says organisations need to independently test these systems before deployment and continue monitoring and addressing vulnerabilities throughout their operational lives.

 

In traditional IT systems, cyber incidents typically result in data loss or service disruption.

 

In autonomous platforms, a compromised drone can be diverted from its mission, an autonomous vehicle can make incorrect decisions, and a robot operating in a hazardous environment can fail at a critical moment.

 

Cyber-physical security protects systems, such as robots, drones, and autonomous vehicles, where software and digital connections control physical actions.

 

It aims to prevent cyberattacks from disrupting operations or causing harm.

 

For decades, electronic products have been certified for safety before entering service.

 

But a platform designed to operate safely can still hide weaknesses in its software, firmware, communications, or vulnerabilities inherited from suppliers.

Verifying security before deployment

 

For organisations that purchase commercial off-the-shelf robots will inherit the software, firmware, third-party components, and supply chains that support them.

 

While vendors perform their own security testing, independent testing can complement vendors’ own checks before a platform enters service.

 

"Today, autonomous platforms and cyber-physical systems need to be evaluated for both safety and security,” says Goh.

 

That means asking practical questions. Where does the platform's data go? How are software updates delivered and validated? How are vulnerabilities identified, monitored, and remediated throughout the lifecycle of the system?

 

Beyond the platform itself, organisations also need visibility into the software supply chain.

 

Understanding what software components are deployed, where they originate from, and whether they contain known vulnerabilities is becoming increasingly important as autonomous platforms become more software driven.

 

Even when proprietary source code is unavailable, security teams can analyse the compiled software and firmware running on a platform to identify potential weaknesses.

 

The risks grow when organisations deploy robots in fleets.

 

"Most organisations don't deploy a single robot. They deploy fleets of autonomous platforms," says Goh. "A vulnerability in one system can potentially be replicated across many systems at once."

 

While tools to monitor IT environments are well-established, securing fleets of autonomous platforms remains an evolving area.

From pre-deployment testing to ongoing monitoring

 

ST Engineering has announced plans to establish a Transparency Centre in Singapore, opening in the first quarter of 2027.

 

The Centre will provide specialised, independent assurance to assess and validate autonomous platforms, from autonomous vehicles to drones and mobile robots, before they are deployed, and to monitor them throughout their operational lifecycle.

 

Drawing on the principles of the NIST Cybersecurity Framework, the centre’s services cover five functions:

 
  1. Identify vulnerabilities and supply-chain risks by analysing deployed binaries, firmware, communications, and infrastructure, without requiring access to proprietary source code.
  2. Protect platforms through trusted hardware, data protection, and secure communications.
  3. Detect threats through penetration testing, red teaming, and continuous monitoring.
  4. Respond to incidents through investigation, containment, and remediation.
  5. Recover through resilience planning and restoration of operations.
 

The Transparency Centre brings these services together to support security from initial testing through to ongoing operations.

 

ST Engineering also designs its own secure integrated circuits and trusted hardware, protecting platforms at the chip level so that only authorised software can run on them.

 

The Transparency Centre will initially focus on sectors where trust, safety, and resilience are critical, including autonomous vehicles, drones, and robotics in public safety, transportation, infrastructure, and industrial operations.

 

"Autonomous platforms, AI, and advanced connectivity are converging rapidly," says Goh.

 

"As these systems become more connected and software-driven, organisations need greater confidence that they are secure, resilient, and operating as intended.”

 

Tan says the aim is to give organisations greater confidence that these systems remain secure and operate as intended as they become more connected and dependent on software.

 

Visitors to the ST Engineering booth at GovWare 2026, held from 13 to 15 October at the Sands Expo and Convention Centre as part of Singapore International Cyber Week, can find out more about the Transparency Centre.