bing track

Artificial Intelligence in Law and Medicine: Data is King

Marty McFly: [Reading the newspaper from 2015] “Within two hours of his arrest, Martin McFly Jr. was tried, convicted and sentenced to fifteen years in the state penitentiary.”? Within two hours?

Doc: The justice system works swiftly in the future now that they’ve abolished all lawyers.

The above scene from the famous Back to the Future trilogy was one of many predictions for the year 2015 (amongst self-lacing shoes, hover boards and flying cars) that it got wrong. At least for the year 2015. Back in the 80s, this would have been the stuff of science fiction and child-like fantasy. But Hollywood’s idea that an iJudge or robo-cop could lay down the law seems to have planted a seed (or microchip) into the future development of artificial intelligence.

No doubt lawyers here in UK will have heard of ROSS, the world’s first artificially intelligent attorney hired by a New York based law firm to deal with bankruptcy matters. Research is also being carried out into computer programs looking to analyse and ‘match’ the decisions of judges using a decision making algorithm. It all sounds rather far-fetched and I suspect many eye brows will be raised at this point, so I’m not going to focus on artificial intelligence in the legal sector on this occasion. All I will say is watch this space.

Artificial intelligence is still in the early stages of development and a lot of work remains to be done, but the best example of its use can be seen in the context of healthcare and medicine. We may not have reached the stage of having a robotic assistant look after us (think Disney’s Big Hero 6) but it is now thought that certain processes carried out by clinicians could be performed more accurately and efficiently using data-driven technology.

For example, artificial intelligence could be used to analyse electronic medical records to deliver high quality data in real time. In the US, clinicians use so called ‘supercomputers’ and systems to identify patients who are at risk of developing kidney failure, cardiac disease or postoperative infections. Doctors will also receive alerts when a particular drug is contraindicated because of a patient’s genetic history and the system can predict which medication the patient will need in future.

One such database available in the US is ‘Modernizing Medicine’, a web based system containing medical information and insights. The idea is to give clinicians access to knowledge shared by other doctors, including the outcome of certain treatments provided to patients with similar conditions.

When you start linking all this pooled data with the wealth of information that can be retrieved from updated publications, a very powerful tool is created. Doctors using this system are better able to deliver treatment plans specifically tailored to their patients.

Another development that is worthy of mention is a glucose meter developed at the Military Institute of Technology (“MIT”) which tests blood sugar levels without breaking the skin. It virtually looks inside the body and measures the level of glucose in the blood. Named ‘iSULIN’, the device is worn on the forearm like a telemetric band (a gadget which measures blood pressure and oxygen saturation) monitoring changing blood flow in peripheral vessels underneath the skin and measuring the absorption of light by tissues. The signal is then transferred to a smartphone app, processed and computed by algorithms which confirm the blood sugar level.

If you think that is impressive, the app goes a step further and offers tips for diabetics on their dieting and can also remind them (with a smartphone alert) that it is either time for a meal or for another dose of medication.

We might as well call these devices healthcare assistants.

Vinad Khosla of Khosla Venture predicts that this sort of technology will replace 80% of what doctors do and enable them to practice medicine more effectively. A lot of what doctors do can be better done by way of data collection and analytics, leaving them able to focus on reaching a more informed and accurate diagnosis.

He emphasises that healthcare is really the “practice of medicine” rather than the “science of medicine” and I am inclined to agree; a trial and error approach is far better suited in the lab than on the front line in Hospitals where patients are being treated.

Ultimately, it’s about acting in the best interests of the patient and that means challenging certain practices of medicine based on tradition that do not have a scientific basis.

When we begin to recognise our own limitations and the limitations of technology, we can come up with a more balanced compromise that will reduce costs and maximise efficiency. Of course machines are not so good at making inferences and understanding the context of notes made by doctors. However artificial intelligence is an extremely powerful tool in crunching data and identifying patterns in a way that humans cannot.

We have generally seen the transition to automation with drones, driverless cars, algorithmic trading, robotic concierges and even robo-monks. Healthcare and medicine will also face change. Whether machines will ever be able to offer appropriate treatment plans seems far away and for now can be left to the imagination of Hollywood.

We need to be prudent in determining how far we invite artificial intelligence into our lives. In the context of healthcare, my view is that it should be limited to the pooling of shared information and analytics, but the clinician should continue to oversee the overall care regime provided to a patient and not solely rely on a system.

By: Magda Zimnicki on in News