Increased productivity through AI?

MTI Viewpoints

Insights shared by industry relative to healthcare and the advancement of medical technology.

Professor Dr Christian Johner is CEO of the Johner Institute. This MTI Viewpoint is from his weekly Executive Corner of the Medical Device Briefings.

“I expect fundamental changes to our medical device ecosystem; not all companies will survive them. The ability to consistently use AI will also determine who will be among the survivors and even the winners. AI will not only be part of the devices but also an internal tool.”

Problem: Finding the right tools

Companies need to find the right tools for their needs. When are general tools such as Large Language Models (LLM) like ChatGPT suitable? When are highly specialized and possibly self-developed AI systems recommended?

Hints for selection

General LLMs have proven themselves in many standard tasks that follow a defined process and can be processed with publicly available information. Examples include creating draft texts, basic data analysis, or developing simple code structures.

Specialized AI tools become relevant when three core factors come together:

1. The task requires access to current, non-public, or highly specialized data sets.
2. In-depth knowledge of a specific domain is needed, e.g., to achieve regulatory conformity.
3. The work processes are complex and rarely standardized.

Example from the medical device industry

LLMs are useful for creating a general product description for a medical device, summarizing regulatory documents, and brainstorming, e.g., in a preliminary hazard analysis.

On the other hand, a specialized tool would be more useful for creating and reviewing regulatory-compliant technical documentation. This tool would not only have access to the latest standards and laws but also to product- and technology-specific rules that have been extracted from the regulations and (!) validated, as well as to reviewed and “labeled” examples.

Further thoughts

Have the courage to try things out! But answer these questions:

• Do you know how to implement the ISO 13485 requirements for computerized systems validation for these tools?
• Have you calculated the return on investment (ROI) and considered the operation and maintenance costs of these systems?
• Have you considered whether you will become an operator (“deployer” as defined by the AI Act), and what requirements you will face?

See more articles from Johner Institute

The post Increased productivity through AI? appeared first on MedTech Intelligence.