In the Netflix sci-fi film Atlas, Jennifer Lopez playing the role of an analyst with a deep skepticism of Artificial Intelligence (AI) is commissioned to track down the renegade humanoid Harlan, who is intent on creating his own version of a better humanity by destructive means. On the perilous mission that takes her to a hostile planet in the Andromeda Galaxy, Atlas suffers a severely fractured tibia that renders her incapable of taking on her adversary, even from the relative safety of her ranger’s pod. Reluctant at first to accept a solution proffered by Smith, the pod’s AI program, she finally begins to bond with Smith, trusting it enough to accept medical treatment for her condition.
Atlas’ remarkably quick and precisely-executed treatment is typical of the role that AI is increasingly playing in the field of medicine, in all spheres – predictive, diagnostic, operative, post-operative and recovery.
AI and the healthcare industry
The global healthcare market currently stands at USD 9.8 trillion with a projection to reach USD 11.9 trillion by the year 2025 and USD 15.1 trillion by 2030. It is already one of the leading sectors in most economies in the world and one of the highest employers.
Since it first began to be applied in the early 1970s, when the MYCIN (1) program was deployed to determine blood infections, AI has played an increasing role in healthcare. Over the years, it has been instrumental in producing faster data collection and processing, assisting in precise surgical procedures, research, and analytics of patient health records.
The global AI healthcare market size was estimated at USD 19.27 billion in 2023 but is projected to grow at a staggering 38.5% CAGR by 2030, driven by the demand for faster, more accurate, and better patient treatments and outcomes.
Cybersecurity in healthcare
Keeping pace with the AI market size is the global healthcare cybersecurity market size (2) which was estimated at USD 17.3 billion in 2023, with a projection to grow at a CAGR of 18.5% by 2030.
Cybercrime in the healthcare industry has come a long way since the 1989 case involving trojanized floppy disks distributed at a WHO Conference in Sweden. The 2014 case of Distributed Denial of Service (DDoS) attack on the Boston Children’s Hospital and the 2017 case of the WannaCry ransomware hack which exploited systems across 150 countries however, remain the benchmark cases for cyberattacks on the healthcare industry. COVID-19 brought with it new cases of cyber threats that raised concerns about data privacy.
In the US, the increasing number of attacks is a cause of concern for the US Department of Health and Human Services. In its December 2023 update, the Office for Civil Rights (OCR) noted that cyber events in healthcare had spiked 93% in four years from 2018 to 2022, with a staggering 278% increase in large breaches involving ransomware.
A good attack surface
To understand why AI in healthcare provides such a good attack surface, one must look at the strides that AI has taken in the field. It continues to transform the healthcare and medicine canvas and is redefining doctor-patient relationships.
Wearable technology, for example, is already allowing patients to monitor their vital signs, sleep patterns, and glucose levels in their home environments, while offering alerts to both patient and caregiver, in the event of threshold limits being breached, unusual behavior being recorded or the device developing a defect. Personal insights of the patient are being generated in parallel as AI algorithms convert behavioral patterns to baseline analyses.
AI-trained algorithms are also powering diagnostic imaging and screening to accurately churn voluminous data residing in the Large Language Models (LLM) when analyzing CT scans, MRIs, X-rays, and other medical visuals, and detecting abnormalities. Deep Learning models are already at work in critical care treatment, providing greater accuracy, fewer instances of misdiagnosis, and fewer false positives and negatives than when done by trained technicians.
In the field of operative and post-operative care, procedures requiring a high degree of precision are being facilitated with AI-powered robotics.
With such a wide array of applications and correspondingly attack surfaces, it stands to reason that medical is a goldmine for bad actors.
Cybersecurity concerns
Notwithstanding other concerns like legal issues and resultant liabilities in the event of failure of AI programs, AI poses several threats, that can result in medically unacceptable situations that violate established patient rights to best-in-class treatment. They include:
- Compromise of patient data in the event of a data breach, which is invariably followed by ransom demands
- Corruption/data poisoning (3) of the LLMs (6) that store and analyze medical data, resulting in misdiagnosis due to falsified results and recommendations
- Malfunctioning of software used to implement surgical procedures, thereby impairing operations and surgeries
Given that AI is expected to grow exponentially in the next 5 to 10 years (4), experts predict cybersecurity incidents will register a similar increase in the corresponding period.
The concerns have sparked the debate about the use of Responsible AI that factors ethical principles in AI applications and workflows to reduce risks, uncertainties, and negative outcomes, while at the same time maximizing positive outcomes. Governments the world over have already begun introducing legislation that will make its use mandatory for healthcare establishments.
For a deeper understanding of how collaboration at all levels will contribute to staving off the cybersecurity dangers inherent in AI, read our blog (6) on the subject.
Final words
That AI is here to say is no longer a matter of debate. Yet as to how soon it will assume full-fledged status in all walks of life, is perhaps boldly answered by the time in which the film Atlas is cleverly set. The year 2043, just two decades from now, resonates with the predictions of futurists like Ray Kurzweil (5), who set a similar date for the maturity of AI systems and human intelligence working in tandem.
A new era looms; it’s time to heed the wake-up call that the day of AI is upon us.
In the field of medicine, it would be life-changing.
References
- The Evolution of AI in Healthcare – Xsolis
- Healthcare Cyber Security Market Size & Share Report, 2030 (grandviewresearch.com)
- Data Poisoning and Exactly Why Organizations Need to Take It Seriously – Aurora Systems Consulting Inc. (aurorait.com)
- Artificial intelligence in healthcare: transforming the practice of medicine – PMC (nih.gov)
- The Potential of Artificial Intelligence in Cybersecurity – Aurora Systems Consulting Inc. (aurorait.com)
- Managing the Risks of Generative AI and LLMs Through Technological Advances – Aurora Systems Consulting Inc. (aurorait.com)
