What is Artificial intelligence?

06 June 2017

What does it mean?

Artificial intelligence (AI) is very much a live topic, with much talk of smart machines and robots driving cars and taking over jobs. But there are also implications for cyber security.

AI is broadly speaking the science of making computers or machines intelligent. The technology, which encompasses areas like machine learning and deep learning, enables machines to carry out tasks that require a degree of intelligence, like understanding natural language, recognising faces in photos or being creative.

The concept of AI has been around since the 1950s – mathematician Alan Turing was an early pioneer, setting out his thoughts in a 1951 paper The Imitation Game. Some argue that progress has been slow since, but the pace has picked up and limited forms of AI are now powering machines that can interact with humans as well as learn from their mistakes.

For example, AI is already powering Google’s search engine, Amazon’s shopping recommendations, and Uber’s taxi App. It is even being used to pilot a network of balloons that provide broadband in Peru.

The technology is also behind developments in automation, such as self-driving cars, as well as in data analytics, where it is being used to diagnose cancer in healthcare, fight fraud in banking and understand complex systems like logistics.

Why you should care?

AI is likely to have implications for cyber security and technology risks in the future. On the one hand, an increased use of AI could bring about new risks, while on the other it could become an important tool in the war against cyber crime.

AI is an important driver behind automation. But as more and more tasks and decisions are automated, there is a risk that errors - in data or in an algorithm - could be compounded. In one real case, a pneumonia diagnostic tool wrongly classified asthma sufferers as low risk because it misinterpreted the underlying data.

There are also concerns that AI-enabled hackers could lead to an explosion in complex cyber attacks. According to the Harvard Business Review, AI could result in more sophisticated social engineering, more intelligent computer viruses and an increase in the attack-surface that hackers can target.

AI systems could, in themselves, bring about new cyber security challenges. Data breaches involving intelligent systems are unchartered territory, and could give rise to new privacy risks, revealing new levels of personal information.

The flip side, however, is that AI could play a valuable role in cyber security, helping to detect cyber attacks, to identify system vulnerabilities and create new antivirus software. For example, a version of IBM’s AI-enabled system Watson has been trained to spot malicious threats in a company’s network.

Download Cyber Decoder

For further information, please contact Sarah Stephens, Head of Cyber, Content and New Technology Risks on cyber@jltgroup.com