Natural Language to the rescue

You can’t teach an old dog new tricks but what about old people? My elderly mother, who lives alone, has owned a panic button for many years but just never got into the habit of wearing it regularly. Not even after having a nasty fall, which left her sprawled on the floor for most of the night, did she learn a lesson.

Could natural language interaction provide a better way to monitor the growing number of old people who live alone? I certainly hope so.

Falls in the home are the most common cause of injury among older adults accounting for more than 75% of domestic accidents for people over 75. They are also the most expensive category of injury for healthcare systems and as western populations get older the financial burden will increase.

The most commonly used monitoring device has a wireless panic button, which connects the occupant to an operator via speakerphone when it is pressed — this is the system installed in my mother’s home.

But these devices are inherently flawed as they require the user to remember to wear the panic button and, equally importantly, to be physically and mentally capable of activating it after an injury has occurred.

Another issue, often overlooked, is the stigma that many elderly people feel about having a big plastic pendant hanging round their neck, which effective marks them out as having lost their independence. In the case of my mother, this was undoubtedly the main reason for her poor compliance.

Yet another problem is that these systems depend on a call center, which has to be permanently staffed. As well as the high cost, such systems are basically “dumb”. The call center operators can provide some remote verbal support but if help is needed, their role is limited to alerting a family member or emergency response team.

Researchers into assistive technologies find that, contrary to the popular perceptions, older adults are open to using new technologies to make their life better. In particular, they are very receptive to technologies which are controlled using speech alone.

A pioneering project employing this approach was the Millennium Home work carried out by the Brunel Institute of Bioengineering in the UK.

A conventional house was converted into a smart home using off-the-shelf sensors that indicate if a user may need assistance — a motion sensor detects a sudden movement, or a water sensor detects that a tap has been left running, for example.

Using loudspeakers distributed throughout the home, the system could play back various pre-recorded messages alerting the occupant of an incident or asking the occupant if they needed this help.

To reassure the elderly person, these messages could be recorded by a family member using personalized phrases delivered in a familiar voice. The system would then listen to the occupant’s response using natural language technology to decide whether the user was capable of resolving the incident or whether to call for external help.

The whole idea of the Millennium Home was to “empower” elderly people to help themselves deal with minor incidents — a bath tap left running, for example — so boosting their autonomy and reducing the burden on call centers and emergency services.

The promoters of the Millennium Home deliberately chose natural language interaction for the “man machine interface” because of elderly people’s greater acceptance and familiarity with spoken communication and their often limited computer skills.

The Millennium home initiative is over a decade old and natural language interaction technology has improved dramatically since then, thanks to the big advances made commercial by pioneers such as Artificial Solutions.

Today, people give spoken instructions to their smartphone and the device is capable of some amazing tricks — such as putting up a Google map showing the nearest pizzeria if you ask it for pizza.

But elderly people are unlikely to own smartphones. So why is there so little interest in packaging and adapting commercial off-the-shelf natural language technology to help the elderly?

Thankfully, there are some people working into this vital yet neglected area, such as the Intelligent Assistive Technology and Systems Lab (IATSL) at the University of Toronto.

The IATLS is involved in many projects that use technology to allow elderly adults retain their autonomy at home. One such project, on “personal robots” is frankly far-fetched, but its project on intelligent fall detection systems would save lots of lives and money if such systems were widely deployed.

The current decade is likely to see enormous advances in areas such as natural language interaction, pervasive computing, artificial intelligence and sensor networks. The convergence of these technologies opens up the exciting prospect of smarter homes that alleviate the burden of dependency and improve the quality of life of our elderly.

Andy, who lives with his family in the UK, is Chief Marketing & Strategy Officer at Artificial Solutions. A regular speaker at industry conferences and events, Andy delivers insight on the rise of AI, the challenges businesses face and the future of intelligent conversational applications.

Leave a Reply

Your email address will not be published. Required fields are marked *