As we release more of our medical tasks to artificial intelligence, we’ll see technology handle things way beyond what humans can do. Artificial intelligence is designed to handle repetitive tasks that can easily be identified using algorithms. Like diagnosing illnesses. Feed in symptoms and AI can maneuver through patterns and connect dots until it comes up with the answer.
Over time, technology will continue to improve on itself. It will learn from its mistakes. This is referred to as machine learning. It improves the results it gives by learning as it handles more tasks.
Within machine learning is another subset called deep learning. This takes machine learning to an entirely different level. Instead of humans taking part in the training, deep learning is built to allow software to train itself. It creates a multilayered neural network filled with information to continue the training process.
The upside to this is very quickly, technology can find even the tiniest triggers that impact a solution. A human brain can only hold so much data. It can only learn so much in a lifetime. But a neural network is limitless. It can always learn, continuously training itself for more. This is a good thing. But your mind can also start to wander and think about all the bad implications as well.
With deep learning, there is no transparency. AI works within, creating its own rules as it goes. If it turns down a wrong path, it’s difficult to find where it went wrong because the learning process keeps going, building an intricate web.
This opens it up to all sorts of security risks, even risks we as humans would never fall for. Have you read about AI mistakenly calling out desert photographs as child pornography?
What if hackers figure out how to penetrate this deep learning mode and feed it wrong information? It’s quite easy to see how it could potentially build different pathways to different conclusions.
Or what about data poisoning? If you feed technology enough “fake” data, eventually it begins to believe it’s true. As it uses algorithms to learn and recommend, it can assume large influxes of data are reality. And that can send entire societies down wayward paths.
You only have to sit through a handful of sci-fi movies to let your imagination run wild.
Is there anything you can do now as we move towards this reality?
Audit - No matter how big or small your database, no matter what kind of information you store, you have to know what your assets are, and the best way to protect them. This is something you should be doing regularly, and changing things as technology advances.
Attack - When is the last time you attacked your own data to find out what happens? This can help identify your weaknesses and protect yourself from potential breaches.
Accountability - Become accountable for every decision you make. Understand the “why.” Why are you taking every step you do? Why are you investing in new technology? Why are you trusting third-party platforms?
The future is guaranteed to be filled with new technology and new security threats. There will never be a time where either go away. Are you prepared to embrace both with open arms?
For IT Strategy, Cloud Conversion, or Help Desk Services reach out to us at Silver Linings Technology 360-450-4759.
