Skip to content

OT Cybersecurity

Sabotage and Human Error - Underestimated Threats in OT Environments

Insider threats and human error in OT - incidents, statistics, and 12 organizational and technical measures based on IEC 62443-2-1, among other sources.

Michał StępieńMichał Stępień

Published Updated 10 min read

insider threathuman errorOTIEC 62443personnel securitysabotage
Sabotage and Human Error - Underestimated Threats in OT Environments
Contents
  1. Why the human factor dominates in OT
  2. A few numbers
  3. Three faces of insider threat
  4. The Great Blackout of 2003 - when one error stops a continent
  5. IEC 62443-2-1 - requirements for the human factor
  6. 12 organizational and technical measures - checklist
  7. The cost of insider incidents
  8. The scale of the problem - SANS 2025 data
  9. What to do in practice
  10. Where we help
  11. Sources

Turn of January and February 2000. On the outskirts of Sunshine Coast in Australia, residents begin complaining about a stench. Bicycle paths along the creek are covered in foul-smelling sludge. Aquatic life suffers. Raw sewage appears on the grounds of a hotel. For two months, no one can identify the cause - sewage pump failures happen regularly, after all.

It was not until April 2000 that it became clear that a single person was behind at least 46 incidents: Vitek Boden, a former employee of Hunter Watertech, the company that installed the SCADA system at the Maroochy Shire Council treatment plant. Boden, who had been denied employment by the council, drove around the area in a car filled with stolen radio equipment, issuing commands to pump stations. In total, he released 800,000 liters of raw sewage into parks, rivers, and hotel grounds.

The court sentenced Boden to two years in prison. One of the first publicly described attacks on a control system was carried out by an insider, that is, a person whose access was granted by the organization, such as an employee or a contractor.

Why the human factor dominates in OT

OT environments differ from typical IT environments in one important respect: operators have direct, physical access to processes that can cause real harm - chemical spills, power outages, destruction of equipment worth millions. The consequences of every mistake, whether deliberate sabotage or a simple configuration error, cannot be undone with one click.

In OT environments, the SANS ICS/OT 2025 survey (more than 330 respondents) shows that 21.5% of organizations experienced an incident in the past year, and 40% of incidents disrupted operations. Nearly one in five incidents needed more than a month to restore normal operations.

A few numbers

Value What it means Source
81 days Containing an insider incident takes 81 days on average. DTEX, Ponemon Cost of Insider Risks 2025
$17.4M Insider incidents cost an organization $17.4 million a year on average (up from $16.2 million in 2023). DTEX, Ponemon Cost of Insider Risks 2025

Three faces of insider threat

Insider threats are not limited to malicious saboteurs. The Ponemon Institute’s Cost of Insider Risks 2025 report classifies them into three categories:

1. Negligent insider

Most commonly, this is an operator who skips a procedure, uses a shared password, or connects an unverified USB device to an engineering workstation. The causes are routine, fatigue, or lack of awareness.

Example - Davis-Besse, 2003: The SQL Slammer worm entered the Davis-Besse nuclear power plant network in Ohio through a contractor’s computer connected to the internet, bypassing the firewall. For nearly 5 hours, the Safety Parameter Display System (SPDS) was unavailable, and operators had no access to the plant’s safety data. The plant was in a refueling outage at the time and was not generating power. The causes were an operating system patch that had been available for six months and a contractor connection that bypassed the firewall.

2. Malicious insider

Deliberate action: sabotage, data theft, process manipulation. Motivations include revenge after termination, financial gain, or pressure from third parties.

Example - Tesla, 2018: Tesla accused a former employee, Martin Tripp, of writing software that hacked into its production control system and of leaking company secrets to journalists. Tripp described himself as a whistleblower, and the dispute ended in 2020 with a settlement in which he paid Tesla $400,000.

Example - Nevada, 2020: Russian citizen Egor Kriuchkov offered an employee of a large US company in Nevada one million dollars for introducing malware into its network that was meant to steal data and be used for extortion. The employee reported the matter to the employer, who notified the FBI. Kriuchkov pleaded guilty in 2021. This case demonstrates that the boundary between an insider and an external attack often blurs.

3. Compromised account (compromised insider)

An attacker takes over an employee’s login credentials and acts on their behalf. From the perspective of monitoring systems, this is a “normal” user performing “normal” operations, which makes detection difficult.

Example - Oldsmar, 2021: In February 2021, at a water treatment plant in Oldsmar, Florida, the sodium hydroxide dosing setting was changed remotely from 100 to 11,100 ppm, and an operator reversed the change. The Pinellas County sheriff presented this as a break-in. In 2023 the FBI stated that it could not confirm a targeted intrusion. A former city manager pointed to employee error, but neither the FBI nor the employee confirmed this. The cause remains unexplained. According to CISA, poor passwords and an outdated operating system were the likely means of access, and the use of remote desktop software was not confirmed. The conclusion for OT is the same for an intrusion and for a mistake by an authorized user: a remote setpoint change without a second confirmation and without value limits is dangerous. Safeguards must therefore work even when an authorized person makes a mistake, that is, include limits on physically plausible setpoints, two-person approval of critical changes, and alarms on setpoint changes. Details are in the analysis of the Oldsmar incident.

The Great Blackout of 2003 - when one error stops a continent

On August 14, 2003, shortly after 4:10 PM Eastern Time, 55 million people in eight US states and the province of Ontario in Canada lost power. Restoring power took from two hours to four days, and losses in the US were estimated at $4-10 billion.

The blackout was caused by a sequence of human and technical errors:

  1. Neglected infrastructure - FirstEnergy’s high-voltage lines touched overgrown trees in a transmission right-of-way that FirstEnergy had not trimmed adequately
  2. Software bug - a race condition bug in the energy management system disabled alarms at the FirstEnergy control center for over an hour
  3. Lack of situational awareness - operators did not know the grid was overloaded and failed to redistribute the load in time
  4. Cascading failures - without intervention, a local failure spread across the entire northeastern transmission grid

WARNING

The 2003 blackout needed only neglected maintenance, an alarm that did not work, and operators who lacked a complete picture of the situation. In OT, a lack of procedures and awareness can have effects comparable to a deliberate attack.

IEC 62443-2-1 - requirements for the human factor

The 2009 edition of IEC 62443-2-1 defines requirements for a cybersecurity management system (CSMS) in industrial automation environments. The 2024 edition is titled “Security program requirements for IACS asset owners” and has a different structure. The requirements of the 2009 edition concerning the human factor concentrate in a few topics:

Topic (2009 edition) Requirement Objective
Awareness and training The standard requires a cybersecurity training program for all personnel, including contractors, both initial and periodic. Fewer errors result from ignorance.
Personnel access control Privileges are assigned according to the principle of least privilege. An incident has a smaller reach.
Account management The standard requires regular account reviews and removal of unnecessary accounts. A risk assessment decides between an individual account and a shift account, and a shift account requires additional physical measures. User actions can be attributed to individuals.
Personnel vetting Persons are screened before access to critical systems is granted, subject to national law. The risk from insiders decreases.
Change management Changes to system configurations are approved under a formal procedure. Unauthorized modifications do not enter the system.
Incident response The organization has an incident response plan. Insider threat scenarios are an additional practice, and the standard does not require them. The problem is identified and isolated quickly.

The application of IEC 62443 to zones and conduits is described in the article on Defense in Depth in DCS systems.

TIP

IEC 62443-2-1 requires a training program for all personnel, including contractors, and recommends training matched to the tasks of the people who design, maintain, and administer the system. A DCS operator needs different training than an OT network administrator or a maintenance manager.

12 organizational and technical measures - checklist

The twelve measures follow from the requirements of IEC 62443-2-1, from the CISA material on insider threat management teams, and from practices described by the Ponemon Institute. They reduce the risk from the human factor in OT.

People

  • Background vetting - screening candidates before granting access to critical systems, subject to national law
  • Role-tailored training - operators, engineers, and contractors each need a different program
  • Incident reporting culture - an employee who reports an error must not be penalized, otherwise errors stop being reported
  • Offboarding procedure - immediate revocation of access for departing employees (the Boden case at Maroochy Shire shows what happens when this is missing)

Processes

  • Four-eyes principle - critical changes to OT configuration require approval from two people
  • Vendor management - control of contractor remote access, connection logs, time restrictions (Davis-Besse: a contractor connection bypassed the firewall)
  • Access reviews - quarterly verification of who has access to what; removal of stale accounts
  • Insider threat response plan - a separate playbook for incidents involving employees, covering legal and HR aspects

Technology

  • Individual accounts - every user has their own account on HMI stations; a shift account requires a risk-based decision and additional physical measures
  • Privileged activity monitoring - session recording on engineering workstations and operator consoles
  • Segmentation and remote access control - limiting an attacker’s movement between systems in the network (lateral movement) even after perimeter breach. Details are in the ICS remote access guide
  • Behavioral alerts - detecting anomalies in user behavior: logins at unusual times, access to systems outside scope of duties, mass data downloads

The cost of insider incidents

The Ponemon Institute’s 2025 report estimates that the average annual cost of insider incidents is $17.4 million per organization, up from $16.2 million in 2023. A negligence incident costs $676,517 on average, an incident with a malicious insider $715,366, and an incident with credential theft $779,707.

Companies spend an average of $211,021 on containing an incident, but only $37,756 on monitoring. They therefore spend 5.6 times more on containment than on monitoring.

Time also has a price: insider incidents contained within 31 days cost an average of $10.6 million, and those that drag on beyond 91 days cost $18.7 million.

The scale of the problem - SANS 2025 data

The SANS State of ICS/OT Security 2025 report shows that:

  • 21.5% of OT organizations reported a cybersecurity incident in the past year
  • 50% of these incidents were initiated by unauthorized external access
  • Only 14% of respondents feel fully prepared for emerging threats
  • Organizations that involve field technicians in exercises are 1.7x more likely to report high readiness

What to do in practice

Insider threats in OT are reduced by consistent work in three fields that organizations often neglect:

  1. Awareness - people know what can go wrong and what warning signs look like. Training builds competence.

  2. Processes - procedures exist, are current, and are followed. An offboarding procedure that no one executes does not protect against the next Boden.

  3. Technology - monitoring, segmentation, and access control help detect incidents and limit their impact. They work when people do not bypass them.

Where we help

SEQRED supports organizations in building OT security programs: from IEC 62443 audits to testing organizational measures during red teaming.

Sources

  1. Verizon 2025 Data Breach Investigations Report
  2. DTEX - 2025 Ponemon Cost of Insider Risks Global Report (press release) (accessed: )
  3. DTEX - 2025 Cost of Insider Risks takeaways (accessed: )
  4. StationX - Insider Threat Statistics (secondary source, cites Ponemon 2025) (accessed: )
  5. SANS 2025 State of ICS/OT Security
  6. SANS - 2025 State of ICS/OT Security report (blog) (accessed: )
  7. CISA - Assembling a Multi-Disciplinary Insider Threat Management Team (2026)
  8. CISA AA21-042A - Compromise of U.S. Water Treatment Facility (accessed: )
  9. CyberScoop - Oldsmar water treatment incident (10.04.2023) (accessed: )
  10. Water Online - Former official claims Oldsmar hack was operator error (29.03.2023) (accessed: )
  11. Control Engineering - Maroochy Shire insider attack
  12. INL CyOTE - Case Study Maroochy Water Services (accessed: )
  13. Control Engineering - Davis-Besse SQL Slammer (accessed: )
  14. MIT - Cybersafety Analysis of the Maroochy Shire Sewage Spill
  15. Wikipedia - Northeast blackout of 2003 (accessed: )
  16. U.S.-Canada Power System Outage Task Force - Final Report on the August 14, 2003 Blackout (accessed: )
  17. U.S. Attorney's Office, District of Nevada - Russian national pleads guilty (Kriuchkov) (accessed: )
  18. Fortune - Ex-Tesla employee settlement (01.12.2020) (accessed: )
  19. ISA/IEC 62443 Series of Standards

Need help in this area?

Our experts will help you assess the risk and plan next steps.

Talk to an expert

We'll discuss scope, methodology, and timeline.

Book a call