In the 2020s, as the technology behind artificial intelligence matures to business as usual, the conversation around it will shift to understanding its place in society. Still, today the comment ‘robots will take our jobs’ is a common concern of everyday participants in the economy. The conversation must turn to how artificial intelligence (AI) is fundamentally changing the workforce, calling out many cognitive tasks, like administrative, legal and accounting tasks, as ones that can be easily automated. It was the industrial revolution which slowly disrupted the prevalence of repetitive physical tasks in our economy, in this, the fourth industrial revolution, we will see these tasks traditionally seen as those uninterruptable by machines. In the 1990s, only thirty years ago, we talked about evolving our economies to ‘service based’ economies. What happens when we can automate the services?
A website aptly titled, ‘Will Robots Take My Job?’ predicts that the jobs at highest risk of automation (99% likelihood) are:
- Data entry keyers
- Cargo and freight agents
- Mathematical technicians
- Library technicians
- Insurance underwriters
- Watch repairers
- Title Examiners, abstractors, and searchers
- Tax preparers
- Sewers, hand
- Photographic process workers and processing machine operators
- New accounts clerks
If you’re currently studying or working in any of these areas, you might start to worry about your future. Even if you’re not, you should be aware of what the disappearance of some of these professions might mean for our most vulnerable citizens. More on that later.
But if you research further afield, it seems no job is safe. Earlier this year, I read an article about AI taking over the jobs of drivers (for example, chauffeurs, truck drivers and even taxis and Uber) in the next 15-25 years. And a recent article from CNBC talks about whole industries where there may be job losses to artificial intelligence: retail, business and logistics, automotives, and marketing and advertising. This is profoundly important, 34,000 Australians are employed as truck drivers, even more in the broader transportation industries.
It surprises me every time I hear that artificial intelligence (AI) is competing for human’s jobs. I strongly believe that AI should be viewed as complementary to a human’s role. AI Is very strong at task specific roles, but it cannot reason, hypothesise, or create new knowledge. Further, AI today struggles to operate effectively outside its training distribution, this ability to generalise to new environments and situations rapidly is still a distinctly human advantage The future is therefore likely a partnership between AI and humans, each playing to their strengths.
Why don’t we like AI?
Maybe it’s the word ‘artificial’, meaning ‘fake’ which turns humans off artificial intelligence. After all, ‘fake’ and ‘artificial’ have negative connotations, are deemed unnatural or contrived and aren’t viewed as authentic, honest or transparent. Or maybe it’s the decades of stories we’ve been told by science fiction writers brandishing visions of the AI future—I bet none of them expected it to be so boring.
It could be the word ‘intelligence’ that humans don’t like? After all, it is rare individual who enjoys being outsmarted and certainly no one likes being the dumbest person (or thing) in a room.
Perhaps it’s just that the majority of humans are still averse to change, and we need to be coaxed into and warmed up to new concepts, such as artificial intelligence. It’s difficult to trust something that is unnatural, smart and unfamiliar.
But imagine the possibilities available to humans when AI acts as a sidekick able to achieve cognitive feats completely out of reach of humans—but working in hand with deep subject matter expertise.
How AI is helping advance cancer research
Take the Immunotherapy Outcome Prediction (IOP) project currently being led by artificial intelligence and machine learning consultancy, Max Kelsen as an example. The specialist Machine Learning and AI organisation is using deep learning to better predict how patients will respond to immunotherapy treatments for cancer.
With the help of genomiQa, QIMR Berghofer, BGI, and QLD Health Metro North Hospital and Health Service, Max Kelsen is integrating AI and whole genome sequencing into cancer research and clinical practice.
IOP aims to develop a prediction model that will indicate the likely effectiveness of cancer treatments such as immunotherapy based on a patient’s unique genomic profile. If successful, the project could help extend patients’ lives, increase their quality of life, and eliminate hundreds of millions of dollars spent on successful and unsuccessful treatment costs.
Why can’t a human do this work?
At 3.2 billion base-pairs per whole genome sequence, each whole genome is almost 300GB. The key to predicting patient treatment outcomes lies in finding and interpreting the patterns and genes of significance in the genomes of patients who have responded best to previous treatments. Understanding correlations at this scale is simply impossible for humans, or simple statistics, but is the type of problem AI thrives on.
Artificial intelligence can find patterns in cancer patient data at a scale, speed and with an accuracy that just isn’t possible for a single research team. The more data, both in volume and variety, we can use to identify trends, the sooner we can develop diagnostic tools which ensure each individual patient is administered the optimal treatment for them and their cancer. AI and Machine Learning has a profound role to play in improving human health, our impact on the environment, and many other challenges of our time. Our work in cancer treatment is just a small part of that