AI tools will change how interactions occur between different health system agents and we need to ensure that the focus of these tools are outcomes such as the empowerment of patients and–by alleviating the workload on healthcare workers–a reduction in provider burnout.

The transition from human-to-human interaction in conventional settings to machine-human interfaces in tele-health is predicated–in large part–on the development of computer-based understanding of regular conversation. Language comprehension is required for even the most basic of computer-to-human interaction.

There is a need for seamless communication and connection across different elements of healthcare delivery. Medical personnel cannot always be present, which creates a need for tele-support / tele-care: healthcare delivered remotely. AI can help address this need. It can do this by enabling the intelligent information and communication environment in which clinicians may interact. AI can also help maintain detailed, real-time updates on the progress via a well-synced network of (virtual) patient-management.

Such technology has an integral role to play in on-line intervention. For it allows for toggling–as needed–between human-guided, patient-guided, and computer-guided approaches. The value of synchronous computer-generated dialogue has led to a broad range of health applications. Automated conversational interaction offers a plethora of opportunities to supplement–and in some cases replace–human carer tasks. These may include:

* Reminders; motivational messages (re: medication, nutrition, and exercise)

* Routine condition checks; health maintenance (based on personalized monitoring)

* Quick answers to health inquiries; provision of targeted health information

* Community involvement; means to address social isolation

* Acting as an intermediary between multiple carers or service agencies

The nature and complexity of a virtual assistant (qua conversational agent) can vary considerably. For simple tasks (e.g. a message or cue), a voice or visual text communication is often adequate. A “smart” system can also convey a response of minimum complexity–such as a confirmation / reminder / acknowledgement, informing the recipient of new / relevant information.

Chatbots capable of audio inputs and outputs can be tremendously helpful; as the interface emulates that of regular human interaction. The simpler technology may have strictly circumscribed capabilities for verbal articulation / language comprehension. Such systems work from constrained conversational models–capable of recognizing and constructing only a limited range of phrases for a pre-specified purpose. These solutions are better suited for interactions where the context of the situation and the user are simple and clearly established.

The more rudimentary AI mechanisms for these agents currently typically rule-based using decision tree logical constructs rather than machine-learning technology. Less sophisticated chatbots seek to mimic conversation rather than understanding it–as with providing travel directions or performing search engine tasks. However, the possibilities are now profound. The level performance rises with a deeper understanding by the AI; and a concomitant increase in the ability to engage in wider (global) data accumulation.

Virtual assistants that can provide a viable alternative to traditional healthcare delivery models drastically improve accessibility to vital online clinical information. And if issues arising from a patient’s past interactions and/or medical history need to be considered in making CONVERSATIONAL choices, a personalized model of the individual’s context (personal preferences / concerns, psychical issues, etc.) can be used.

The first chatbot called ELIZA was able to establish a conversation with human beings and mimic their conversational models. It operated primarily by rephrasing input sentences–such that they matched a set of pre-defined rules. Another evolving technology is voice recognition–made commercially popular by Google Home, Microsoft’s Cortana, Amazon’s Alexa, and Apple’s Siri. Such technologies make it possible to have a service that can attend to a patient and caregiver in need of help at any given time. As we know, with formidable advances in language-learning software and in AI, the capabilities of such technology are increasing at a rapid pace.

Hybrid technologies where both humans and chatbots interact with patients will ensure that our limited medical resources can be allocated in the maximally efficient way.