The integration of digital technologies into physiotherapy practice is accelerating rapidly. 

Artificial intelligence (AI), wearable sensors, telehealth platforms and automated clinical tools are no longer emerging concepts. 

Rather, they are increasingly embedded within everyday care. 

These tools offer significant opportunities to enhance clinical decision-making, improve access and support patient engagement. 

However, they also introduce new layers of complexity, risk and ethical responsibility. 

At the centre of this transformation is a critical question: how do we ensure safe, ethical and accountable care in an increasingly automated environment? 

Increasing complexity and the burden of proof 

As digital systems become more sophisticated, they also become more difficult to fully understand. AI-driven tools, in particular, often rely on complex models that operate beyond the intuitive grasp of the clinician. 

While these systems may produce highly precise outputs, the reasoning behind those outputs is not always transparent. 

This creates an important shift in practice as clinicians are required to interpret and justify decisions that may be influenced by systems they do not entirely control or comprehend. 

For example, a physiotherapist might use an AI-supported movement analysis platform to assess gait following injury. 

The system may highlight subtle deviations and recommend a management plan and a set of progressive rehabilitation exercises. 

While this appears helpful, the clinician must ask: how reliable is this output? 

Under what conditions has the system been validated? 

Does it account for this patient’s specific characteristics?

It is important to understand that the presence of technology does not reduce responsibility but rather, in many cases, increases it. 

The clinician must not only remain confident in what decision is made but also be able to justify it clinically. 

Evaluating digital tools for clinical use 

Safe digital practice depends on the ability to critically evaluate the tools being used. 

This involves more than simply understanding how to operate a system; it requires an appreciation of its intended use, its evidence base and its limitations. 

A clinician introducing a wearable device to monitor patient activity, for instance, should consider how accurately that device measures movement in real-world conditions. 

A system validated in controlled laboratory settings may perform differently in a patient’s home environment. 

Factors such as fatigue, clothing or sensor placement can influence the quality of the data generated. 

There are also regulatory considerations. 

In Australia, tools intended for clinical use may require oversight from bodies such as the Therapeutic Goods Administration. 

However, regulatory approval does not eliminate clinical responsibility. 

It simply indicates that a tool meets minimum safety and quality standards. 

Ultimately, the physiotherapist must be able to determine whether a tool is appropriate for the specific patient and context in which it is being used. 

The ‘black box’ problem and clinical oversight 

A defining feature of many AI systems is their lack of transparency. 

These ‘black box’ models generate outputs without clearly explaining how those outputs were derived. 

This presents a direct challenge to clinical accountability. 

When a system recommends a particular course of action, the clinician must be able to understand and justify that recommendation. 

In practice, clinicians must treat AI outputs as one component of the decision-making process rather than as definitive answers. 

This requires an active, reflective approach. 

Outputs should be compared against established evidence, considered within the broader clinical picture and tested against the clinician’s own reasoning. 

For example, if an AI system flags a patient as being at high risk of re-injury but the clinician’s assessment suggests otherwise, the discrepancy must be explored.

Is the system detecting something subtle that warrants further investigation? 

Or is it generating a false positive based on limitations in its training data? 

Maintaining this level of oversight ensures that technology enhances, rather than undermines, clinical judgement. 

Bias, equity and the risk of inequitable care 

Professor Trevor Russell says responsibility in digital practice must be shared across multiple stakeholders.

Digital systems are inherently shaped by the data used to develop them. 

These datasets are rarely fully representative of all patient populations, which introduces the risk of bias. 

As a result, systems may perform more accurately for some groups than for others. 

In physiotherapy, this has practical implications. 

A tool designed to assess movement patterns may be less accurate for individuals with atypical biomechanics or disabilities. 

Similarly, predictive models may not account for cultural, social or environmental factors that influence health outcomes. 

Clinicians must therefore remain attentive to the limitations of digital tools and consider whether outputs are appropriate for the individual patient. 

This requires not only technical awareness but also professional judgement. 

Equity also extends beyond algorithmic performance. 

Access to digital tools is not uniform and reliance on technology may inadvertently disadvantage patients who lack access, resources or digital literacy. 

Ensuring equitable care in a digital environment requires thoughtful implementation and flexibility in how services are delivered. 

Privacy, security and the management of health data 

The use of digital tools in physiotherapy inevitably involves the handling of sensitive health information. 

Whether through telehealth platforms, wearable devices or AI-driven documentation systems, patient data is being collected, stored and transmitted in new ways. 

This introduces significant responsibilities in terms of privacy and security. 

Clinicians must ensure that systems are used appropriately, that data is protected and that processes are in place to respond to potential breaches. 

These responsibilities are not purely technical; they are central to maintaining patient trust. 

For patients, the implications of digital data use are not always obvious. 

A simple telehealth consultation may involve multiple layers of data processing, including recording, transcription and storage in cloud-based systems. 

Without clear communication, patients may not fully understand how their information is being used. 

Informed consent in the digital context 

Informed consent remains a cornerstone of ethical practice but it becomes more complex in the context of digital health. 

Patients must be provided with a clear understanding of how technology is being used in their care. 

This includes explaining the role of AI or digital tools, outlining their limitations and describing how patient data will be managed. 

For example, if a consultation is being transcribed using AI, patients should be explicitly informed and given the opportunity to opt out. 

Importantly, this information must be communicated in plain language. 

The technical complexity of digital systems should not become a barrier to patient understanding. 

Clinicians must translate these concepts into terms that are meaningful and accessible, enabling patients to make informed decisions. 

Consent should be explicit, documented and revisited over time. 

Patients should also be reassured that they retain the right to withdraw consent or request alternative approaches to care. 

Shared responsibility and the ‘safety net’ 

While clinicians remain accountable for patient care, responsibility in digital practice is shared across multiple stakeholders. 

Developers are responsible for designing systems that are accurate and secure. 

Healthcare organisations must provide appropriate governance, training and oversight. 

Regulators, including the Therapeutic Goods Administration, ensure that tools meet safety standards before they are implemented. 

Patients also play an important role, as informed participants in decisions about their care. 

Together, these groups form a ‘shared safety net’. 

However, within this system, the clinician remains the central decisionmaker. 

This reinforces the need for critical engagement with digital tools and a clear understanding of their role in practice. 

Training and maintaining clinical capability 

The increasing use of digital technologies requires physiotherapists to develop new skills. 

This includes understanding how algorithms function, recognising their limitations and integrating them effectively into clinical workflows.

At the same time, there is a risk of over-reliance on technology. 

If clinicians defer too readily to automated systems, there is potential for deskilling, particularly in areas such as clinical reasoning and patient communication. 

Maintaining strong foundational skills remains essential. 

Technology should be viewed as an adjunct to practice, a tool that enhances capability rather than replaces it. 

Maintaining accountability in a digital future 

Digital technologies offer significant opportunities to improve physiotherapy practice. 

They can enhance assessment, support decision-making and expand access to care. 

However, they also introduce new ethical and clinical challenges that must be carefully managed. 

The central message is clear: technology does not replace accountability. 

The clinician remains responsible for the care they provide, regardless of the tools they use. 

As physiotherapy continues to evolve, the ability to engage critically with digital systems will become increasingly important. 

By maintaining strong clinical oversight, prioritising transparency and consent, and focusing on patient-centred care, physiotherapists can navigate this transition responsibly. 

In an age of automation, it is not the technology itself that determines the quality of care but how it is used. 

>> Professor Trevor Russell is the director of RECOVER Injury Research Centre at the University of Queensland. His research focuses on the use of digital technologies for the remote delivery of health services.