6.0 Artificial Intelligence

Framework structure

Primary domain 6, artificial intelligence, is shown in this section.

AI refers to the ability of machines to mimic human intelligence or behavioural patterns. In practice this often refers to the automation of various activities that involve tasks like finding patterns in data, and making predictions.

Capability statement - I understand that Artificial Intelligence (AI) is an umbrella term used to define digital technologies capable of performing tasks commonly thought to require human intelligence. I am aware AI is common in modern technology and can list uses of AI outside healthcare (for example, voice recognition, recommender systems, self-driving cars, image and video processing)

Archetypes:

  • Shaper
  • Driver
  • Creator
  • Embedder
  • User

Capability statement - I can provide examples of AI systems used in healthcare and understand their potential benefits and risks (for example, imaging diagnostics and decision support tools)

Archetypes:

  • Shaper 
  • Driver
  • Creator
  • Embedder
  • User

Capability statement - I am aware that “machine learning” is a subset of AI and is an umbrella term used to refer to techniques that allow computers to learn from examples/data without being explicitly programmed with step-by-step instructions

Archetypes:

  • Shaper
  • Driver
  • Creator
  • Embedder
  • User

6.1 Machine learning and natural language processing

Machine learning (ML) is a sub-set of AI. Machine learning and natural language processing offer models and techniques for making predictions and classifications with large amounts of data. They can do this without the need for the explicit programming of step by step instructions, learning instead from the data provided.

Capability statement - I understand that machine learning algorithms require large quantities of data to learn from, and must be trained and evaluated using independent sub-sets of the available data

Archetypes:

  • Shaper
  • Driver
  • Creator
  • Embedder
  • User

Capability statement - I am aware of some of the common uses for Natural Language Processing (NLP) methods and text mining within and outside of healthcare (for example, chat bots, speech/virtual assistants, dictation of clinical notes, processing electronic patient records)

Archetypes:

  • Shaper
  • Driver
  • Creator
  • Embedder
  • User

6.2 Using and implementing AI systems

AI systems often have to be tailored for specific situations and uses. In the healthcare environment it is less likely that processes will be entirely automated. This often involves humans working with AI systems (human in the loop) where AI systems automate some (usually routine or repetitive) tasks to free up time for specialists to focus on other areas. This sees a closer working relationship between human experts and AI systems.

Capability statement - I am able to use Artificial Intelligence (AI) systems confidently to assist me to improve task efficiency while maintaining quality and safety

Archetypes:

  • User

Capability statement - I know how to respond if an AI system fails or is inaccessible and can initiate an alternative process to maintain effective service provision

Archetypes:

  • Embedder
  • User

Capability statement - I understand the importance of sharing learning following failures of an AI system, to improve systems and practice

Archetypes:

  • Creator
  • Embedder
  • User

6.3 Evaluating AI systems

Once AI systems have been implemented they need to be evaluated to ensure they are safe and fit for purpose. Algorithmic drift can occur prompting the need for continual readjustment. The system also has to be evaluated in the local context to ensure it is configured correctly and working optimally for the chosen goals.

Capability statement - I can explain the difference between internal validation, external validation, local evaluation and prospective clinical evaluation of Artificial Intelligence (AI) technology and their relevance to clinical performance. I am aware of recommended standards for validation of different types of AI technologies used in healthcare (such as the National Institute for Health and Care Excellent (NICE) evidence standards framework for the digital health technologies)

Archetypes:

  • Driver
  • Creator
  • Embedder

Capability statement - When evaluating an AI system for use in my professional workflow, I can compare its performance against the expected standards in my professional area of practice

Archetypes:

  • Driver
  • Embedder
  • User

6.4 Robotics

Robotics builds on many of the AI technologies and represents the intersections of these areas. Robots are being used for routine cleaning and disinfecting procedures, logistical processes, high risk areas (for example, highly infectious individuals) and surgical tasks through to providing companions for the elderly. Robotic technology takes on various forms and applications and is set to increase in use in the near future as the technology improves and the cost reduces.

Capability statement - I am aware of the use of robotic technology for healthcare and can cite examples of the use of robots for health, medical and social care (for example, social companion robots, surgical robots, room disinfectant robots, nanotechnology)

Archetypes:

  • Shaper
  • Driver
  • Creator
  • Embedder
  • User

Page last reviewed: 14 February 2023
Next review due: 20 February 2024