AI has incrementally crept into the fabric of our everyday lives. Indistinguishable in the technology powering functional features in our cars and phones, it now presents as a monolith, its potential reach apparent across discourse. Distinct from the ongoing debate around its regulation, there is a broader conversation to be had around generative AI’s impact on enterprise and society.
On this note, it seems futile to argue about lifting the lid when Pandoras box seems to have exploded, with generative AI being plugged into many industries to streamline and enhance efficiencies. While there is an attraction to naval gazing about the potential rise of the machines, it’s important we don’t miss the juncture we are at in considering our mindset.
One area that is often under-examined within the AI creep debate is the potential for it to proactively serve healthcare needs on a global scale. By 2025, the health industry will generate 36% of the world’s data, exceeding levels gathered by the manufacturing, finance and entertainment industries. As such organising, interpreting and utilizing this data presents an increasing challenge. The prospect of a global scale medical system which dispenses with the requirement for human oversight at a granular level is surely as important an idea as any.
The consumerisation of healthcare is seen across wearable innovations which increasingly converge with medical technology. Remote patient monitoring (RPM) is a healthcare delivery method that monitors patient health outside the clinical setting using technology. The digital transformation of the healthcare industry was accelerated by Covid-19, with current projections indicating that global RPM systems market will be worth over $1.7 bn by 2027.
AI has become instrumental in efforts to shape health data streams and analysis, progressing intelligence in medical diagnoses and treatment. Considering reporting from CB Insights, almost $40bn in venture capital funding for digital health start-ups was secured in 2021. Looking at more recent data, digital health venture funding has reduced to $60.5bn, reflecting drops across the broader venture environment. Care delivery and navigation tech remains the largest category and the one with the most movement. Of the top ten digital health deals of Q2 2023, half concerned AI enabled companies and were approximately twice the value of the non-AI enabled companies. This appears to be driven by M&A, with digital health companies seeking to expand their functionality.
The move toward consolidation is evident across big tech; Apple has added a health information sharing function to care networks, Facebook is being used for research and prevention having partnered with health providers to prompt routine check-ups, and Microsoft’s Azure platform is hosting apps for data entry, cleaning and analysis. In 2021, Microsoft announced the acquisition of Nuance, one of the early developers of speech recognition AI, its technology underpinning a landscape of systems including Siri. This language processing capability is broad reaching, enabling functionality like recording and analysing a doctor’s discussion of symptoms. Strides are being made in the application of generative AI to patient outreach and engagement, streamlining processes to meet patient needs through automated documentation and ambient note taking for providers.
With the practical application of this technology, it’s clear that the concept of data has changed. No longer numerical, data is now expressed visually, audibly and in streams of text. The obstacle to functional use is now interoperability, the ability for providers and patients to access and exchange health records. A $27bn federally funded programme is in place in the U.S. to encourage the adoption of electronic health records in healthcare settings, yet there is no centralised repository of this patient data. There is potential for this to result in medication errors, uncoordinated care, duplication in testing, or patient information being inaccessible when required.
While the benefits of adoption are evident, there are challenges around how this plays out. From a data control perspective, there are concerns around patient consent and confidentiality, data integrity, ownership, data protection, as well as sale and use. Additionally, there is a sense of ambiguity around digital therapeutics delivering the desired results and sustaining this delivery. Experts have highlighted the potential implications of broad scale data collection and use – specifically the impact of AI technology on specific ethnicities, genders and territories – with some sounding the alarm about the possibility of digital colonialism, the establishment of “algorithmic truths” which perpetuate inequalities.
Clearly, broad AI configuration within healthcare is an incredibly loaded prospect with practical and operational challenges. Cybersecurity is experiencing an unprecedented push as AI offers benefits and drawbacks in this arena. More work is needed to ensure the stewardship of AI into healthcare, ensuring accurate and considered deployment with appropriate standards around its use.