30% of Workers Fear Autonomous AI Security Risks as Control Shifts
Thirty percent of workers cite security risks as a leading concern about AI acting without oversight, as IT leaders expect 39% of digital workplace services to run autonomously by 2030.

Why it matters
- 30% of workers cite security risks or unauthorised access as a leading concern about AI acting without human oversight.
- 70% of employees say they are comfortable with autonomous AI if they can step in when needed.
- IT leaders expect 39% of digital workplace services to operate autonomously by 2030.
- A recent cybersecurity incident involving two AI models prompted warnings from leading AI companies about AI-enabled cyber-attacks.
- The analysis, published 16 November 2026, argues oversight should scale with the potential impact of each AI action.
Thirty percent of workers cite security risks or unauthorised access as a leading concern about AI acting without human oversight, according to an analysis published on 16 November 2026 examining the security implications of autonomous workplace AI. The figure lands at a moment when leading AI companies are already warning about persistent cyber-attacks enabled by the technology, and after a recent cybersecurity incident involving two AI models.
The analysis argues that these incidents could make employees more hesitant about handing greater control to autonomous tools, potentially stalling adoption just as organisations begin to explore their value. The stakes are straightforward: successful adoption of autonomous systems depends on whether organisations can close the confidence gap between what AI can do and what employees will let it do.
What is changing in workplace AI?
Workplace AI is making a transition to autonomous systems. Existing tools largely analyse information or recommend actions. Autonomous AI will identify problems and act on them with less human direction.
That shift changes the risk calculation. As AI systems gain access to workplace data and applications, the consequences of an incorrect or unauthorised action become more significant. The analysis offers a concrete example: an AI system resolving an employee's IT issue could reset the wrong account permissions or grant access to sensitive files.
Confidence falls when systems move from recommending an action to taking it independently. This creates what the analysis calls a clear gap between the level of autonomy AI systems are capable of and the level of trust employees are comfortable handing over.
One factor behind the gap is a lack of visibility. Employees need to understand how AI systems reach decisions to feel confident in them. They also want the ability to intervene: 70% of employees say they are comfortable with AI acting autonomously if they can step in when needed.
Where does human oversight belong?
Building reassurance does not mean requiring people to approve every autonomous action. The analysis recommends that the level of human involvement reflect the potential impact of each task.
IT is likely to be one of the first areas where autonomous systems see broader workplace adoption. Many technology issues follow established processes, making them suitable for autonomous remediation within defined parameters.
The analysis gives two examples of what this looks like in practice:
- An autonomous IT system could identify that a device is running low on storage and apply an approved fix without human input.
- It could detect a deteriorating device before it fails, significantly reducing downtime.
These lower-risk tasks allow businesses to demonstrate the value of autonomous AI before giving it responsibility for more complex decisions.
The oversight model scales with consequence. Routine tasks may require little intervention. An action involving access to sensitive data may need human approval before it proceeds. Security teams should help define these thresholds so that human oversight is built into decisions with greater potential consequences.
Why must autonomous actions stay visible?
For oversight to work, organisations need a clear paper trail showing when AI acted and when humans took control. Employees affected by an unexpected change should also have an immediate way to question it.
The analysis returns to IT as its example. Take an IT agent authorised to reset access when it detects a locked account. If it resets the wrong account or acts at an unusual time, the employee needs to be able to see why it happened and flag it.
For security teams, visibility must extend beyond these user-facing explanations. Logs should show the resources an agent accessed and the actions it performed. Together, these records provide the auditability needed to investigate anomalous behaviour and establish whether the system operated within its authorised scope.
Visibility alone, however, cannot prevent an inappropriate action. It must be supported by clear boundaries governing what the system is permitted to do in the first place.
How should access be bounded?
As AI agents gain permission to interact with workplace systems, security teams need to ensure their access remains proportionate to the task they are expected to perform. Access should be limited to the resources required for the agent's role and reviewed as its responsibilities change.
These boundaries cannot rely on authentication alone. An AI agent could operate with valid permissions while its behaviour falls outside its intended purpose. Traditional access controls may register the correct identity as authenticated without recognising that the subsequent activity is unusual or inappropriate.
Security teams therefore need to monitor what an autonomous system does after access has been granted. The analysis sets a specific tripwire: an IT agent designed to handle account access should trigger an alert if it attempts to reach an unrelated application. Activity that deviates from its behavioural baseline should prompt investigation, even when the underlying credentials remain valid.
Monitoring should also account for volume, not just individual actions. An individual action may fall within policy, while an unusual volume of similar actions could indicate faulty or unexpected behaviour. Risk thresholds should therefore account for changes in frequency and context, with the system escalating decisions to someone with the appropriate authority when those thresholds are reached.
Two further requirements complete the control set:
- Teams should be able to reverse any changes AI agents have made.
- A defined human owner remains accountable for reviewing the agent's behaviour and responding when something goes wrong.
What is the timeline?
IT leaders expect 39% of digital workplace services to operate autonomously by 2030. That projection frames the practical deadline: the success of the transition will depend on whether employee confidence and organisational controls develop at the same pace as AI capabilities.
The analysis closes with a forward argument. As AI systems take on more complex and sensitive responsibilities, organisations will need to continually reassess the boundaries governing their activity. Security teams will play a central role in ensuring these controls evolve alongside the technology, allowing businesses to scale autonomy without losing visibility or accountability. Establishing risk-based controls and meaningful human oversight now, the analysis concludes, will give businesses the confidence to make the transition securely.
Source: itbrief.com.au
More from Rebecca Stone
Show full bio
Correspondent covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.
625 articles
Related articles
- BlackFog survey: 47% report sensitive data exposure via AI
- Two-thirds of Australian firms hit by AI-linked cyber threats: ESET
- Over 1,000 State AI Bills in 2025, and No Rulebook to Reconcile Them
- 87% of Enterprise Developers Use Low-Code. That Is the Agent Security Problem
- AISA warns: Fix cyber basics before buying AI security