Speech Recognition in Health
VoiceSignals #18 — Musings on Voice tech news
😷 We’ve mentioned it before, voice can help medical staff detect illness like heart conditions or respiratory problems. For voice to be produced it needs the help of our lungs, our vocal cords, the muscles surrounding them, and their health can play a role in how voice is produced or what it sounds like. Scientists are looking more intensely now as to how AI can detect indicators of ‘coronavirus sounds’ in our speech. Professor Cecilia Mascolo, an expert in mobile health at data analysis at the UK University of Cambridge, is collecting sounds from COVID-19 patients to train her machine learning models to recognize infections. Created as part of a project called EAR, she hopes it might eventually lead to new ways of diagnosing respiratory diseases and help in the global fight against coronavirus. It’s all part of a wider scientific approach to diagnosing SARS-CoV-2 in time and space, enabling policymakers and healthcare workers to track and mitigate the outbreak of COVID-19.
💊 When will AI truly change healthcare? While many have been touting the coming AI revolution for healthcare the solutions we’re seeing are still missing a key factor. Patient care. EHR systems with integrated speech recognition, that automatically fill in the fields, sound great but in practice doctors are staying glued to the screen in fear of generated mistakes, which could prove lethal. That drives away their attention from the actual patient. Imagine what affect that would have in an ER theater. Harjinder Sandhu, CEO of Saykara, a company working with physicians to streamline their workflow, explains how human scribes have freed physicians to focus on their patients. He wants to build an autonomous AI scribe, and although he comprehends the complexity, he argues that “AI systems augmented by human power, are filling in the gaps of AI competency and allowing these systems to succeed while incrementally chipping away at the goal of making these systems fully autonomous. These systems are beginning to do more than simply relieve doctors of the burden of documentation, though that is obviously important. The real transformative impact will be from capturing a comprehensive set of data about a patient journey in a structured and consistent fashion and putting that into the medical records, thereby building a base for all other AI applications to come”.
🚸Online education and COVID-19 proved that e-learning still has many challenges to resolve before it becomes a norm. Young children are more sensitive and complicated in the way they communicate. An Irish based startup, SoapBox Labs raised $6.3M for its speech recognition technology that focuses specifically on understanding children between 3 and 12 years old. Their app started out as an education solution for childhood illiteracy and now is able to understand and capture significant behavioral, physical, and even environmental differences between children and those who are older. By narrowing their focus on young children, and how they communicate, gives them an advantage that most speech recognition software solutions out there don’t have; children’s speech. Something very important to understand when developing toys with speech recognition or educational software that needs to treat children’s’ learning with sensitivity and empathy.
As always, stay safe whether in or out of isolation 🙂.
In today’s edition:
🏦 AI, Banking and debt repayment efforts
💸 How AI can help Banks plan for a surge in non-performing loans
— Vicki Kolovou
What we read online…
How AI can Help Banks Plan for a Surge in Non-Performing Loans (NPLs)
The number of non-performing loans (NPLs) during the 2009 Financial Crisis doubled from pre-crisis levels. Through the end of 2019, the US had $95bn in NPLs and Europe €636bn for a total of 0.91% and 3.1% respectively. If those numbers double during this pandemic crisis, as they did in 2009, it will represent a significant strain on the banks. At the same time, the current methods of debt collection in contact centers are inefficient and largely driven by human operators and analysis. Behavioral profile pairing and predictive models that evaluate customers in real-time can help to transform this process with better profiling and prioritization for the restructuring of these NPLs. Read more >
How banking benefits from AI and voice technology
Rana Gujral, CEO at Behavioral Signals, writes on how voice technology is leading the way in improving customer experience and augmenting human customer service agents to be better at their jobs. In addition to processing millions of transactions and customer interactions every day, banks are now also able to evaluate customer voice data in real-time to identify their behavioral patterns in order to best match customers with agents to achieve desired results. This technology will be essential to not only speed up the process to improve real-time efficacy, but also provide better support and customer satisfaction when faced with such challenges as debt payments and loans. When the right customer is matched with the right customer-service agent, not only will the interaction be optimized to produce the desired result on both ends, it will also lead to higher satisfaction from both client and employee. Read here >
The Power of Voice Technology: By Everyone, For Everyone
Our upbringing and our life experiences shape our view — of how the world is, and how it should be. That in turn, shapes the solutions that we create and the challenges we choose to tackle. With emerging technologies such as voice and artificial intelligence, we have the opportunity to re-shape our future and create a more equitable society for all. We can all be active participants in writing the next chapter of humanity that reflects our values and priorities — for a world that is as diverse and vibrant as ever… Listen here >
Voice App Developers Survey
62% of voice app developers reported in January 2020 that building custom voice apps for businesses provides a clear path to monetization for the industry. That figure is up from just 49% in July 2019. However, optimism around voice apps designed with monetization baked in such as for product purchases, subscriptions, or advertising fell precipitously during that same period. Whereas 49% saw voice app products as having a clear path to monetization in 2019, by early 2020 that figure had declined to 30%. Read here >
