Using artificial intelligence (AI) in healthcare
Artificial intelligence can be a game changer for how we offer healthcare. AI is mentioned as one of the tools to be expanded on in the NHS Long Term Plan, which has been set up to ready the NHS for the future.
What is AI?
Artificial Intelligence, often shortened to AI, is an artificial form of human intelligence. AI learns and solves problems like a human does. But because it can process a lot more data, the error margin often is a lot slimmer as it takes away human error.
The difference between AI and any old software system is the capability to adapt and improvise, not simply follow a step-by-step guide, but learn along the way.
Examples of AI are face recognition and voice-to-text software. These systems have to think and process to construct a specific outcome.
Why use AI in healthcare?
There’s a great number of reasons why healthcare can benefit from artificial intelligence. A few examples are:
- AI can steer better health in people. The smart technology can help people change into or maintain a healthy lifestyle. We already see how apps can help us manage our health.
- AI brings more information to light. With artificial intelligence people and therefore clinicians can get a better sense of what the daily lifestyle of a person looks like and therefore can much more accurately offer the right kind of intervention, whether medical or non-medical.
- AI has the potential to diagnose more accurately as well. As technology is able to store vast amounts of information – much more information than an individual is able to store – the dataset for AI is much larger, leading to more accurate diagnoses.
- AI for the elderly. Loneliness is an increasing health risk for the elderly, in particular those that do not have a large support system. AI is offering a way for the elderly to have social interactions accessible.
What are the current issues to implement AI in healthcare?
There are a lot of arguments around in support of AI. There are few that do not see its benefits for patients and clinicians alike.
However, the implementation of AI in healthcare is a more difficult matter. Trishan Panch, Heather Mattie and Leo Anthony Celi in their 2019 article ‘The “inconvenient truth” about AI in healthcare’ argue that the solutions on offer are not ‘executable at the frontlines of clinical practice’.
They state that sustainable change is not possible via AI systems as the healthcare system is fragmented as individual organisations purchase hardware and software separate from each other. Secondly, most organisations do not have the required data infrastructure to train these AI systems to the requirements of the organisation’s specific locality.
Despite Panch, Mattie and Celi stating the market is not ready for the implementation of AI, they do offer a glimmer of hope. They say to realise AI’s potential certain issues need addressing, namely ‘who owns health data, who is responsible for it, and who can use it?’. If this data ownership can be centralised then there is a possibility for AI to blossom in healthcare. Without it, AI in healthcare will simply remain a possibility.