AI in AgeTech, Part 6: Balancing AI to Eliminate Falls with ZIBRIO

By Rick Robinson posted 11-27-2023 03:04 PM


In part six of our series looking at how AI is impacting the AgeTech space, we talk with Dr. Katey Forth, CEO of ZIBRIO, whose product emerged from the U.S. space program and wrangles machine learning to help keep people from falling, which by the way is the leading cause for injury in adults over 65.

The ZIBRIO scale takes measurements to predict if you are going to fall in the next year, is that right? Please break it down for us.

When you stand on ZIBRIO’s Stability scale, the scale is measuring the pressure under your feet. There is no such thing as standing still, and the Stability scale is measuring the tiny movements and corrections your body is making to balance. Using ZIBRIO’s proprietary AI algorithm called Briocore®, the scale detects when your body is in control, and when it is having moments of micro-failure of postural control. The output is a number from 1-10, and this score is predictive of whether you will fall down in the next 12 months.


What elements and types of AI are employed to achieve this measurement?

The Stability scale uses machine learning in combination with our novel, punctuated equilibrium model of control to identify your dynamic postural stability and when you exhibit subtle patterns of instability.  This AI technique creates a highly accurate and insightful score that predicts future falls 2-5 times better than previous clinical tools.   With such an objective, predictive score, ZIBRIO enables true fall prevention care, rather than the status quo of waiting for someone to fall before giving them help.

This AI technique creates a highly accurate and insightful score that predicts future falls 2-5 times better than previous clinical tools.


Who developed the AI, or are these more accurately stated as algorithms? 

The patented algorithm was developed by ZIBRIO’s co-founders, myself and Dr. Erez Lieberman Aiden, while working in the U.S. space program to solve balance and falling for astronauts in lunar and Martian explorations and their re-adaptation to Earth’s gravity environment. We both witnessed the consequences of falling from our grandmothers, so it was clear how impactful this AI would be for the health of older adults (on Earth). ZIBRIO was awarded runner-up for NASA Invention of the Year 2023. 


Are you continuing to develop the AI function and technology? 

The current algorithm has been validated for adults in clinical and real-world testing. However, as we work with specific patient populations there is an opportunity to fine tune our technology for that population; for example, patients undergoing chemotherapy, or the Parkinson's’ population. Also, we continue to develop the supportive tech outside the scale, allowing physicians to monitor patients remotely, streamline how the stability score is connected and informs the care plan and the EHR. 


Talk a little about your feelings on the latest iteration of AI

For this question I thought I would use an LLM to find the answer, and this illustrates my feelings. Chat GPT gave me:

“Large Language Models and advancements in generative art had showcased significant progress, raising ethical questions about AI's impact on language understanding, bias, and the creative process

The answer is decent, lacks some depth, nuance, analysis, etc, but makes me seem like I know what I am talking about. 


That's funny... So can LLMs help?

Absolutely. Naturally, we need to pay attention to what it’s been trained on, and be aware of potential biases, and there is also the danger of just being swamped by low quality information. But, if you view it as an idea generator rather than using the words, concepts, and content verbatim, then you can extract value that can improve productivity. I think it is easier to spot those inaccuracies within the art than the LLM.  For the generative art, leaving copyright issues aside, this tool can democratize producing images.

Again, bias is important - has the model been trained on user-centric design or universal design principles, or are we just going to perpetuate the design mistakes of the past? At ZIBRIO, we invest a lot in the human interaction parts of our product - from ease of use, to the colors and symbols and language we use to explain things. I don’t think generative art is up to scratch yet, but I’m sure it’ll get there eventually.


Do you see any use for generative AI in your product or future products?

Absolutely. The opportunity to provide hyper personalization and support for our older users is where we are most excited. 


More broadly, where do you see this “new” AI going, and in particular how might it be applied to new products helping people 50 and over?

My view is that, when used well, this “new” AI is empowering. So, it can be argued that the greatest value of these tools is for those who have less access, capability, or resources. Thus, the tools can counter some of the limitations we develop as we age. 


What concerns you – if anything – about what OpenAI, Google, Meta and other open source AIs are creating? 

It’s a bit like the Wild West at the moment. I’m concerned about a complete free for all - with the ability of bad actors like bot farms flooding the systems and changing people’s understanding - we’ve seen it in the political landscape already. But it has the capability of affecting every aspect of people’s lives, so there will need to be a consensus in how it is managed, what the rules are and then they’ll need to be enforced.

You can find Part 1 in the series here, Part 2 here, Part 3 here, Part 4 here and Part 5 here.