Advances in machine learning and the advent of 5G have given rise to ambient intelligence — physical spaces and devices that sense and learn from the behaviour of humans.
With the global data science market set to be worth nearly $26bn by 2027, real-time insights will be a key future value proposition. Ambient intelligence has the unique ability to capture sensor data, sift it for valuable attributes and patterns, and provide actionable information in real-time.
In healthcare, virtual assistants that use machine learning to streamline clinical consultations are already growing in popularity.
The ambient experience lessens the laborious task of documentation, enabling healthcare professionals to focus on the patient experience. M*Modal combines speech recognition and artificial intelligence to accurately transcribe the physician-patient interaction, perform risk assessments and suggest or correct care recommendations.
As infrared sensors have become cheaper and easier to source, their use in high-risk care-giving environments is increasing. The Stanford experts spearheading the research describe ambient devices that can scan and analyse the everyday movements of clinicians and patients using infrared detection, based on “how long it takes the invisible rays to bounce back to the source.”
The team is also investigating activity recognition algorithms, which can mine sensor input for patterns in behaviour and flag changes. For instance, if clinicians were to show signs of fatigue or fail to wash their hands before a patient interaction, the ambient sensor’s decision-making function would kick in.
In assisted living settings, the same technology could be used to raise red flags if it detects changes in a resident’s normal, healthy behaviour. Based on its sensor data, if an older person was then deemed to be at risk of an accident or showing signs of illness, this could trigger an alert to local caregivers or emergency services. With this unobtrusive layer of protection, older people could live independently for longer.
Home automation is a fast growing trend. Many smart cars, appliances, thermostats and health trackers already rely on ambient intelligence to improve their efficiency. With 5G poised to significantly decrease “latency” — the lag between a network and a connected device — this sector is set to grow even further.
Ever-active technology embedded into the home will raise inevitable concerns over transparency and privacy. Fears that the information could be used to predict and influence consumer behaviour will need to be addressed early on. The reaction to a leaked test feature for Google’s Nest Hub, which would have enabled the device to use its built-in ultrasound capabilities to sense nearby users, will surely act as a precursor for future debates.
Though researchers see hurdles ahead in regulation and public trust, they believe that these can be overcome through a collaborative approach — and that ambient intelligence’s “extraordinary” potential benefits will ultimately win out.
“Preliminary results from hospitals and daily living spaces confirm the richness of information gained through ambient sensing. This extraordinary opportunity … requires computer scientists, clinicians and medical researchers to work closely with experts from law, ethics and public policy to create trustworthy ambient intelligence systems.”
