Building a Future-Ready Pharmacy: Artificial Intelligence & Cybersecurity
In 2025, being a “Future-Ready Pharmacy” includes two core elements – an AI strategy and a holistic cybersecurity plan. In terms of AI, nearly every new product and service released in the last 12 months touted some form of AI integration. However, the reality behind these claims varies widely, from simple rebranding of existing features and incremental enhancements to genuinely transformative innovations. Healthcare systems are increasingly becoming the frequent targets of cybercrime given the sensitive personal, financial, and medical data they store, which are highly lucrative on the black market. Additionally, cybercriminals understand the urgency of restoring functionality to prevent disruptions to patient care. They know that many healthcare systems operate with outdated or insufficiently secured technologies, making them easier targets than other industries. In 2025 and beyond, healthcare leaders will continue to be challenged by cyber threats and tasked with identifying which products truly deliver on AI’s promise and, ultimately, adopt and implement technology that advances the organization’s mission.
Artificial Intelligence
Artificial intelligence (AI) refers to machines and software designed to perform tasks that typically require human intelligence, a concept dating back to 1956 when the term ‘artificial intelligence’ was first coined. Pharmacy has been technology-centric for decades, utilizing contemporary systems and applications to provide safe, effective, and economical care. Examples include dispensing automation, clinical decision support, predictive analytics for inventory management, drug diversion monitoring, and many others. AI now includes generative models that can complete tasks like learning, reasoning, decision-making, natural language processing, and perception to generate new data with many potential healthcare applications.
Generative Pre-trained Transformer (GPT)
When you ask OpenAI’s ChatGPT to define a GPT, you get this response: ‘A GPT (which stands for Generative Pre-trained Transformer) is a type of artificial intelligence model designed for natural language processing (NLP) tasks.’ Developed by OpenAI, GPT models are capable of understanding and generating human-like text. They are based on a neural network architecture called a transformer, which is particularly good at handling sequential data, like language, by capturing complex patterns, relationships, and context within large amounts of text data.
Open AI launched its first GPT model in 2018, which was subsequently followed by models from Google DeepMind (Gemini), Meta (Llama), Anthropic (Claude), and Apple (Apple Intelligence). The explosion of GPT models has led to ‘GPT’ being used as a verb, similar to how ‘Google’ became synonymous with web searching. In addition to these tech giants, healthcare technology companies are integrating these models into software to improve all aspects of healthcare outcomes (e.g., financial, operational, clinical, and humanistic). Table 1 below depicts the potential application of existing AI models within healthcare.
| Model | Primary Use | Applications |
|---|---|---|
| Large Language Models (LLMs) | Text-based data, clinical interactions | Clinical documentation, patient chatbots, summarization |
| Generative Adversarial Networks (GANs) | Image generation and enhancement | Synthetic images, MRI reconstruction, data augmentation |
| Variational Autoencoders (VAEs) | Data structuring & anomaly detection | Anomaly detection in imaging, genomics, drug discovery |
| Flow-Based Models (FBMs) | Synthetic data & density estimation | Synthetic patient data, risk prediction |
| Diffusion Models | High-fidelity data generation | Medical image reconstruction, drug design |
| Transformer-Based Multimodal Models | Integrated data analysis | Clinical decision support, patient monitoring, radiology reporting |
“AI-ification” – “AI washing”
Recent AI breakthroughs and widespread media coverage turned AI into an ambiguous and overused buzzword. The term ‘AI washing’ or ‘AI-ification’ refers to marketing products or services as AI, AI-powered, or AI-ready to appear more innovative or advanced, even when the AI involved consists of simple algorithms, basic data analytics, rule-based automation, or sometimes no AI at all. Companies and individuals engage in “AI washing” to attract investment, generate hype, or capitalize on AI’s popularity in branding.
Healthcare decision-makers will benefit from developing frameworks and governance structures that protect the organization from “AI washing” and ensure the ethical deployment of generative AI, GPTs, and other AI technologies.
Cybersecurity and Downtime Preparedness
Ransomware is malicious software (malware) that renders data and systems inaccessible until a ransom is paid. Healthcare organizations and associated business partners have increasingly become targets of ransomware attacks. Clinical, operational, and financial activities within a pharmacy are highly dependent on technology, which plays a critical role in every phase of the medication use process and beyond (Table 2). In general, the technologies used are reliable; any outages typically impact small operations for short durations. However, there are times when outages are widespread and/or prolonged, caused by ransomware attacks or other precipitating factors such as system failures or natural disasters. Examples include prolonged outages of electronic medical records, disruptions in payment processing systems, and failures in automated dispensing cabinets, with durations ranging from days to months. Given leaders’ day-to-day challenges, downtime preparedness, training, and drilling are often insufficient or non-existent.
Additionally, some pharmacies have had autonomy for technical support outside of information services, which could be vulnerable to ransomware attacks if the systems do not have the requisite defenses and monitoring capabilities to detect and mitigate cyberattacks. Given the patient care and financial stakes of technology downtimes, pharmacy leaders must not only develop downtime plans but also conduct regular training and drills to ensure readiness. It is not a question of whether downtime will occur; instead, it is a question of when and for how long.
| Purchasing | Ordering – Prescribing | Dispensing | Administration | Monitoring | Billing | |
|---|---|---|---|---|---|---|
| Electronic Health Record (EHR) | X | X | X | X | X | |
| Retail Pharmacy Software | X | X | X | |||
| Automated Dispensing System (ADS) | X | X | ||||
| Pneumatic Tube System | X | |||||
| Payment Processing Switch | X | |||||
| Dispensing Robots (Retail and Inpatient) | X | |||||
| Carousels | X | |||||
| Billing processors (Retail and Inpatient) | X | |||||
| Wholesaler Purchasing Platforms | X | |||||
| Inventory Management Systems | X | |||||
| E-prescribing platforms | X | |||||
| Sterile Compounding Platforms | X | |||||
| Point of Sale Systems | X | |||||
| Automated Tablet Packaging Machines | X |
Calls to Action
- Ensure that organizational governance and frameworks are in place to thoroughly vet AI claims and oversee the safe and ethical adoption of generative artificial intelligence.
- Determine when and how pharmacists should be engaged in decisions where technology impacts medication use.
- Partner with your organization’s cybersecurity team to perform a gap analysis of vulnerabilities in pharmacy data and systems.
- Conduct a gap analysis of downtime preparedness and develop a comprehensive plan to address identified gaps.
Looking for more information or assistance for your organization? Reach out to Visante today!
Subject Matter Experts: Joe Lassiter & Phil Brummond