The implications of recent advances in Artificial Intelligence (AI) have spurred heated debate globally. As science fiction starts to become reality, AI products are slowly infiltrating homes and workplaces. This is raising concerns about the potential detrimental effects of AI on the job market, or even about the dangers of an AI singularity, where sentient robots take over the world and destroy humans.
While these are all valid discussions, I believe that the focus of AI should not be just on cool home gadgets or on process optimisation and automation. Instead, AI can be used to fundamentally rethink how we solve the world’s problems.
AI has the potential to greatly improve things like healthcare, education, poverty and security. AI machines can do some very beneficial things already today that humans will simply never be able to. If we leverage that to augment what humans do well, AI could positively impact society, business, and culture on the order of magnitude of the internet itself.
I call this using AI to scale the human mind, not replace it. The human brain is the most elegant computer in existence. We process millions of sensory inputs automatically and constantly, allowing us to learn and respond to our environment. But the human brain only contains about 300 million pattern processors that are responsible for human thought. What if we could complement all of our amazing ideas with not just more data, but also orders of magnitude more data processing capability? Imagine how we would rethink every single problem that exists today.
Even with today’s primitive forms of AI, there is enough technology out there to start doing exactly this. The examples below draw from a variety of industries to illustrate the magnitude of social impact possible when we couple AI with human skill and ingenuity.
1. Precision Medicine
AI is driving the adoption and implementation of precision medicine: an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person. Think of it as a type of medical personalisation. For example, around 25,000 people in the US are diagnosed with brain tumors every year. Traditionally, they might all be given the same course of treatment to see what might work in a one-size-fits-all approach. Precision medicine will allow doctors and researchers to predict more accurately which treatment and prevention strategies for a particular disease will work in which groups of people.
Many of the answers lie in the vast amount of medical data already collected. Ayasdi uses AI algorithms like deep learning to enable doctors and hospitals to better analyse their data. Through their work, medical practitioners have been able to identify previously unknown diabetes sub-types that could lead to better understanding of therapies that could work better for certain types of patients. Enlitic and IBM are using similar AI algorithms but to detect tumors in radiology scans more accurately and efficiently, and even potentially accelerate finding a cure for cancer.
There were around 707 million cybersecurity breaches in 2015, and 554 million in just the first half of 2016. The impact of just a few of these attacks, such as foreign governments potentially biasing US presidential elections, is truly scary.
Security teams struggle today to work through the increasing number of alerts generated by traditional tools. The self-learning and automation capabilities enabled by AI can increase effectiveness and reduce costs, keeping us much safer from terrorism or even smaller scale identity theft.
AI-based solutions already in the market can be more proactive and can preempt attacks in the pre-execution state by identifying patterns and anomalies associated with malicious content. Secureworks uses the predictive capabilities of AI for advanced threat detection on a global scale. SiftScience, Cylance, and Deep Instinct are using it for fraud preventions and for endpoint security, like smartphones and laptops. These technologies will dramatically expand the scope and scale of security professionals and allow them to detect threats hopefully well before they actually attack.
3. Precision Farming
The world's population is expected to increase significantly over the next three decades, but our capacity for food production will struggle to keep pace. AI is driving efficiency in our current farming methods to increase production and reduce wastage without adversely affecting the environment.
Systems such as John Deere's AutoTrac enable huge machines to plant crops in a far more uniform and accurate way and can reduce overlap in agricultural processes such as tilling, planting and fertilising, which in turn reduces the use of chemicals and increases productivity.
Cainthus, a machine vision company, has another approach. Using deep learning, it has created a facial recognition system than can identify individual cows by their facial features in just six seconds, enabling huge herds to be monitored with minimal human involvement. Soon, they will be able to detect early signs of lameness in a cow based on its body shape, and alert the farmer accordingly.
As sensors proliferate on farms and drones capture real-time images of the condition of vast amounts of farmland, AI machines will be able to help farmers foresee what their crops and farms are going to need potentially over a year in advance, giving them more time to react to adverse conditions.
AI can be applied to many more problems and markets. In fact, it should be thought of as a fundamentally new approach to every problem. Those decisions will be made by humans who want to change and improve the world, and who now can scale their minds to address ever-expanding frontiers.