Will Artificial Intelligence Take Over The World?

Artificial intelligence is moving fast. According to a report by Gartner, the capabilities of AI technologies across the board are changing fast, with a higher-than-usual number of innovations in the ‘hype cycle’ (the cycle that all emerging technologies follow as they develop from brand-new innovations to established applications). 

Here’s an exploration of some of the latest advancements in artificial intelligence, alongside some of the concerns these advancements have ignited regarding risks to our cyber security.  

Recent Developments In AI Technology

Of the recent developments in artificial intelligence, perhaps those that are most relevant right now are those expected to reach the end of the hype cycle in the next 2-5 years. These are as follows:

1. Composite AI

As the name suggests, composite AI is quite literally the combination of multiple AI techniques. The types of technology implemented through composite AI techniques are completely dependent on the problem the technology has been adopted to solve. However, some of the techniques that crop up almost every time are machine learning, deep learning, contextual analysis, and knowledge graphs.

The applications of composite AI are almost unlimited. The approach taken when applying multiple AI technologies enables more “human” decision-making processes which are applicable to a huge range of industries: from finance to customer experience. 

2. Generative AI

Generative AI is, quite simply, the artificial generation of content without human input. Whilst it is predicted that this technology has reached less than 1% of its target audience so far, there are examples of generative AI cropping up across the internet. Deepfakes; AI copywriters; and automated coding are just three recent applications of this technology. 

3. Augmented Intelligence

Also known as human-centred AI, augmented intelligence is the term used to describe humans working in partnership with AI technology. It is often used when either AI or human resources have reached full capacity and is generally considered a less risky adoption of artificial intelligence than fully automated processes. A key example of augmented intelligence are virtual assistants like Alexa and Siri - the core principle of such technology is designed to enhance, rather than replace human intelligence. 

4. Synthetic Data

Synthetic data, or artificially generated data, is a kind of data that is not collected in the real world but is synthesised based on knowledge picked up by an AI model. For example, a synthetic data technology may use a provided set of existing real-world data as a basis to generate a huge amount of additional data on the same subject. Its applications are enormous, as it can save organisations time and money in almost any data collection endeavour. However, this technology is early in its emergence and is not especially reliable in its reflection of real-world data right now. 

5. AI governance

AI governance, as you might expect, is a practice through which organisations implementing AI technologies hold themselves accountable for its associated risks. Governments across the globe, including in the UK, US, and Japan, are beginning to implement rules and requirements that specify how AI technology can be adopted, including the level to which they are researched and developed before they can be fully implemented. In the case of AI, governance is extremely important because the risks associated with less well-established technologies are especially high. 

The Risks Of AI Development

The U.S. Department of Commerce's National Institute of Standards and Technology (NIST) recently raised concerns over the risks posed by artificial intelligence. The organisation is planning to draft an Artificial Intelligence Risk Management Framework (AI RMF), a specially devised document aimed at those developing and utilising AI technology. Use of the framework will not be mandatory but will be actively encouraged as a means to improve trustworthiness and reduce the risks associated with the technologies. 

So what are these supposed risks? Here are a few of the risks associated with AI:

• Job losses due to automation of existing human job roles

• Privacy violations initiated by deepfakes and other invasive AI technologies

• Data biases in synthesised data

• Weapons automation (a pretty self-explanatory one, we think!)

• Threats to cyber security 

Cyber Security 

Let’s focus on the cyber security risk for a second. AI technology proposes a specific risk to our collective cyber security because many developments in AI technologies are (both knowingly and unknowingly) making it considerably easier for cyber criminals to attack. 

There are many ways artificial intelligence can support criminal activity online, but the most common are mimicking and vulnerability searching. Through machine learning, AI technologies can watch and learn about an individual’s actions online. Once it has gathered enough information, it can then copy these actions, mimicking the individual and often gaining access to protected information. In addition, programmes running through AI can scan the internet for weak spots in online security, detecting even the slightest changes that may indicate a vulnerability. The criminals utilising the technology can then exploit these vulnerabilities and, if successful, launch a full-scale cyber attack. 

The Role Of Quantum

Perhaps even more concerning is the potential for quantum technology to enhance the abilities of artificial intelligence. Whilst AI may hold rather terrifying potential for cyber criminals, it is limited by the speed and capacity of the computer(s) it is run through. As long as this technology is run through ordinary computers, the time it takes to break through most cyber security measures will render it a waste of time for cyber criminals. However, the introduction of quantum computers will change this forever. 

With a quantum computer, much of the work that would take an ordinary computer (or even a supercomputer) hundreds to thousands of years to complete could be completed in seconds. 

Enter Quantum Encryption 

It’s not all doom and gloom - fortunately, there is a cyber security technology capable of securing our online presence against the risks of both AI and quantum computing: quantum encryption. 

Platform-as-a-Service solution, QuantumCloud™ by Arqit is a simple and effective solution designed specifically to solve the problems of legacy encryption techniques and provides a solid line of defence against quantum attacks.

Another solution that also implements quantum encryption is from Qrypt. The basis of its solution is to make everlasting encryption accessible to all by reimagining how data is encrypted, transmitted and stored. 

Artificial Intelligence has already shown itself to be a revolutionary aspect of technology - but in order to stay one step ahead, security needs to be just as revolutionary!