AI Workplace Surveillance: Guide for Tasmanian Employers

A new Australian report released on 20 October 2026 recommends that worker monitoring should be “fair and reasonable” under the Fair Work Act. That recommendation is not a new law. For Tasmanian employers, the practical response is to audit any app, camera, tracking tool or AI system that collects worker data, define a legitimate purpose, minimise what is collected, explain the system clearly and keep people—not an algorithm—responsible for consequential decisions.

The University of Technology Sydney Human Technology Institute (HTI) report, Drawing the Line: Tech-enabled surveillance of Australian workers, calls for federal reform after consultation with employers, workers, unions, legal experts and government stakeholders. It proposes a dedicated Fair Work Commission complaints pathway, clearer regulatory guidance and room for enterprise agreements and modern awards to add workplace-specific protections.

This article explains what was proposed, what has not changed, and the sensible checks a Hobart or Tasmanian business can make now. It is general information, not legal advice.

What happened on 20 October 2026?

HTI published its final report on technology-enabled worker surveillance and called on the Australian Government to amend the Fair Work Act. Its central recommendation is a requirement that monitoring be fair and reasonable, supported by factors that consider the employer’s purpose, the effect on workers and whether a less intrusive approach could achieve the same outcome.

The report describes technology-enabled surveillance broadly. It can include keystroke and screen monitoring, email analysis, GPS tracking, CCTV with AI capabilities, facial recognition, audio recording, productivity scoring, emotion or sentiment analysis, wearables and software that records online activity.

HTI does not argue that all workplace monitoring is harmful. It recognises legitimate purposes such as work health and safety, cyber security, legal compliance and protecting business property. The concern is that a reasonable objective does not automatically justify every collection method or every later use of the data.

What has not changed?

The report is an evidence-based reform proposal, not legislation. It does not create a new Fair Work Act test, a new reporting deadline or an automatic ban on workplace monitoring. Businesses should not rewrite their policies on the assumption that Parliament has already adopted HTI’s recommendations.

Existing obligations still matter. The Office of the Australian Information Commissioner says employers conducting surveillance must follow applicable federal, state or territory laws. It also notes that privacy principles may apply to some records created through monitoring, while the precise position depends on the information, employer and employment context.

The Fair Work Ombudsman’s workplace privacy guide recommends clear policies that identify what information is collected, why it is collected, who can access it and which areas or activities are monitored. Tasmania-specific surveillance, workplace, privacy, anti-discrimination and work health and safety questions may require professional advice.

Why this matters to a Tasmanian small business

Workplace surveillance is not limited to large warehouses or national call centres. Monitoring features can be built into everyday products used by a small business:

• rostering apps that record location or attendance;
• vehicle or delivery GPS systems;
• CCTV with person detection or facial analysis;
• time-tracking software that captures screens or activity;
• collaboration platforms that report presence or engagement;
• customer-service tools that score calls or messages;
• wearables used for safety or productivity; and
• AI tools that rank applicants, flag errors or assess performance.

A Hobart hospitality operator might use cameras for safety and loss prevention, a vehicle-based business may use location tracking, and an office may rely on productivity or meeting software. Each use has different purposes, risks and alternatives. A single sentence in an employment contract is not a substitute for understanding what the technology actually does.

A practical eight-step review for employers

1. Create a monitoring register

List every system that can observe, record, score or infer something about a worker. Include built-in features that have not been deliberately purchased as “surveillance”. Record the supplier, system owner, data collected, collection frequency, storage location, retention period, access permissions and whether automated analysis is used.

Ask managers and IT providers separately. One may know about cameras and access cards while the other knows about device management, logs and software analytics.

2. Write down the legitimate purpose

State the specific operational need for each collection. “Improving performance” is too broad. “Confirming lone-worker arrival at a remote site for safety” is clearer and can be tested against the data being collected.

Do not quietly reuse information for a new purpose because the system makes it possible. Location data collected for dispatch may not be suitable for scoring every break. Call recordings collected for training may not justify an automated disciplinary decision.

3. Test necessity and proportionality

Ask whether the monitoring is reasonably connected to the stated purpose, whether it operates only when needed and whether a less intrusive option is available. A work vehicle may need location data during a delivery shift; continuous tracking outside paid hours would raise a different question.

Collect the minimum information required. If attendance can be confirmed by a site check-in, recording continuous movement may add risk without adding useful evidence.

4. Check the supplier before switching features on

Request a plain-English explanation of how the product works. Important questions include:

• Which raw data and inferred data does the product create?
• Does the supplier use workplace data to train its own models?
• Where is information stored and which subcontractors receive it?
• Can monitoring be limited by time, location, role or device?
• How are errors, bias and false alerts tested?
• Can a worker see and challenge information about them?
• How is data exported or deleted when the service ends?

DanSol’s guide to custom software versus off-the-shelf systems explains related questions about data control, integration, ownership, security and vendor dependency.

5. Explain monitoring before it starts

Tell workers what is being monitored, when it operates, the purpose, the information produced, who sees it, how long it is kept and how it may affect decisions. The explanation should cover actual product behaviour rather than a generic reservation of rights.

Consult workers and their representatives where required or appropriate. Practical feedback can reveal consequences a vendor demonstration misses—for example, a GPS tool that records private travel, a productivity metric that disadvantages a worker with disability, or a camera angle that captures a changing area.

6. Keep a person accountable for decisions

Do not treat an AI score as an objective fact. Monitoring data may be incomplete, inaccurate or detached from context. A low activity score may reflect training, approved leave, customer-facing work away from the device, accessibility needs or a faulty integration.

For decisions about hiring, hours, pay, leave, performance, discipline or termination, identify the responsible human decision-maker. Give them enough information to test the output, consider context and document why the decision is justified independently of the tool.

7. Secure the information

Monitoring can create a concentrated record of locations, communications, images, behaviour and performance. Restrict access to people with a genuine business need, use strong authentication, review administrator accounts and keep an appropriate deletion schedule.

The OAIC’s guidance for commercially available AI products recommends due diligence, human oversight, regular review, transparency and a privacy-by-design approach when personal information is involved. DanSol’s ASD Essentials guide for Hobart businesses provides a companion checklist for access, cloud services, recovery and cyber resilience.

8. Review outcomes, not just settings

Schedule a review after implementation and when the supplier, workplace or intended use changes. Check whether the tool achieved its purpose, whether alerts were accurate, whether certain workers were affected disproportionately and whether the system changed behaviour in unhelpful ways.

HTI warns that crude proxies such as email counts or meeting attendance may not measure useful output or work quality. A system that generates more dashboards but weakens trust, increases workarounds or misdirects management time is not automatically productive.

A simple fair-and-reasonable decision table

Question	Stronger practice	Warning sign
Purpose	A specific safety, security or operational need	Monitoring because the feature exists
Scope	Limited to relevant people, times and data	Continuous or unrelated collection
Transparency	Clear notice, policy and worker consultation	Hidden settings or vague contract wording
Accuracy	Testing, context and a way to correct errors	Scores treated as unquestionable facts
Decisions	Named human reviewer with real authority	Automatic discipline or disadvantage
Security	Restricted access and defined deletion	Indefinite storage or broad supplier access
Review	Evidence that benefits outweigh impacts	“Set and forget” deployment

How to brief a developer or software supplier

If a monitoring feature is part of a new app, portal or business workflow, include governance requirements in the brief before development begins. Define user roles, audit logs, access controls, deletion rules, notices, review pathways and which decisions must never be automated.

DanSol’s app development service for Hobart and Tasmanian businesses begins with users, tasks and technical requirements. For a worker-facing system, those requirements should include privacy, accessibility, security and a clear explanation of how information influences decisions.

Also ensure the system can be changed. A business should not need to rebuild an entire application simply to shorten retention, disable an intrusive feature or give workers access to their records.

Questions to ask at the next management meeting

• Which technologies currently monitor workers, including features embedded in ordinary software?
• Can we explain the purpose and operation of each one in plain English?
• Are we collecting more information, for longer, than the purpose requires?
• Which decisions use monitoring data, and who remains accountable?
• Could a worker see, correct or challenge inaccurate information?
• Have we assessed impacts on remote workers, people with disability and other affected groups?
• Do contracts let the supplier reuse or retain the data?
• Which Australian and Tasmanian laws, awards, agreements and policies apply to our situation?
• When will the system be reviewed or retired?

Frequently asked questions

Has Australia banned AI workplace surveillance?

No. The 20 October 2026 HTI report recommends reform; it is not a ban or a new law. Existing workplace, privacy, surveillance, safety and anti-discrimination requirements still need to be considered.

Is ordinary CCTV considered AI surveillance?

Ordinary CCTV is a form of monitoring, but not every camera uses AI. Features such as facial recognition, object identification, behaviour analysis or automated alerts add different data and decision risks. Review the actual model and configuration.

Can a business monitor company-owned devices?

Ownership of the device does not answer every legal or fairness question. Consider applicable law, the stated purpose, notice, proportionality, information collected, personal use, access and how the records will influence decisions.

Is worker consent enough?

Do not assume a broad consent clause resolves every issue. HTI argues that consent in an employment relationship may not be freely negotiated, while other legal requirements can apply regardless. Obtain situation-specific advice where necessary.

What should a small business do first?

Start with a one-page monitoring register. List every camera, tracker, app and analytics feature, then write the purpose, data, access, retention and decision use for each. Any blank answer is a priority for investigation.

Prepare for scrutiny without pretending the law has changed

The immediate lesson from HTI’s report is not to panic or purchase another compliance product. It is to understand the technology already in use and be able to justify it. Monitoring should solve a defined problem, use no more data than necessary, be transparent to workers, protect information and keep meaningful human review.

Australian policy may change after further government and parliamentary processes. Tasmanian employers can prepare constructively now by documenting current systems, checking contracts and policies, involving workers, and obtaining advice for high-risk uses such as biometrics, audio recording, emotion analysis or automated employment decisions.

Information checked on 20 October 2026 against the UTS Human Technology Institute report, Fair Work Ombudsman guidance and Office of the Australian Information Commissioner resources. This article provides general business information and is not legal, employment or privacy advice.