Accessability Links

The impact of AI and machine learning in 2018

Artificial intelligence (systems that can learn to act autonomously) and machine learning (where computers use data to learn without being explicitly programmed) are rapidly becoming a part of the IT industry. As AI takes hold, there’ll be a demand for IT specialists with skills in algorithm and training methodology selection, data preparation, integration and model creation. Data scientists and developers will end up working together to create these systems. 

Apps will also gain a degree of intelligence as they incorporate AI and machine learning. Of course, there’s a fear that they could replace people in the workplace, but actually what we may see is intelligent apps enhancing human activity, with machine learning automating data preparation.

The Internet of Things is a key buzz-phrase right now, and AI will pave the way for objects to display advanced behaviours that can help them interact in a natural way with humans and the environment. Driverless cars are of course the big news, although following an accident in Arizona in March 2018, it’s evident that there’s a lot of testing still to complete.

Perhaps the biggest immediate IT impact of AI and machine learning will be felt in cyber security. At present, machine learning systems work alongside human security analysts, rather than replacing them, processing data (and thus allowing analysts to work on other tasks). Machine learning is also very useful for organising data, clarifying decision-making for humans. Certainly, visual representations of data are more straightforward to analyse than rows of numbers.

Machine learning is at its best right now in identifying anomalies: behaviour that might represent an attack on a network through malware or hackers, or spam communications. Huge global tech companies including Google and Amazon are using machine learning to identify and remove malware.

But if machine learning can recognise threats, it can also be used by cyber criminals to launch attacks – perhaps even by tricking security systems. And as soon as cyber criminals realise an attack has been thwarted, they will of course learn to change their approach.

Of course, as machine learning evolves, it could in fact trick the attackers, going on the defensive by fooling cyber criminals into searching for files that may appear real but aren’t. Learning from human behaviour is another potential pathway; machine learning could analyse previous human responses to attacks and assess their appropriateness. 

Read our related blog: Blockchain – is it the future?
Add new comment

Meet the team

Back to Top