AI: Unlimited potential or unparalleled risk?

“I visualise a time when we will be to robots what dogs are to humans, and I’m rooting for the machines.” —Claude Shannon, mathematician and computer scientist.

After that slightly bleak and dated opener, it’s time to lift the lid on AI.

In the last few years, AI (artificial intelligence) has become as common a global acronym as VIP, 3D or MVP. But is AI really your most valuable player or could it be a dangerous and unstoppable threat to not just your team, but to the security of your entire organisation? Especially when we’re talking data protection and possible cyber breaches.

AI in brief:

It may be possible that both the positive and negative implications of AI have passed you by, thus far. But that’s unlikely.

According to IBM, AI is “technology that enables computers and machines to simulate human intelligence and problem-solving capabilities”.

It’s become the driving force behind robotics, GPS, autonomous vehicles and content generation tools like ChatGPT. All of which is amazing; operations can be performed, driverless cars can take to the road, and if we really wanted, this 100 per cent human article could have been written in a minute by a tool.
 

AI in business – the basics and the benefits:

Aside from being the power behind the “internet of things” whereby all across a factory machines can talk to one another, manufacture goods then move them about without human intervention, AI is moving fast and furiously through the world of data processing – and access and incorrect use of this type of information and related tools brings with it a set of challenges and potential issues.

For example, when a chatbot answering system may be much more efficient to answer many common customer service questions, where and when do we change this auto response to a real person who can interact and intercept? 

Many people in business are also utilising elements of AI to create emails, spreadsheets, or information brochures for employees and clients. So, what are the key benefits of implementing some form of AI technology in the workplace?

1. Automated workflows: Routine processes can be simplified and completed much quicker. Whether this is running reports, code creation or the aforementioned chatbot responses, there are a host of jobs AI can take out of human hands. 
2. Virtual assistance: This is the occasion when you can trust the AI software to create templates for emails and handle employee and customer communications, usually triggered by voice commands. 
3. Understanding human behaviour: Beyond the creation of the communications that go out to customers, AI through machine learning now has the capability of monitoring and analysing human responses and behaviour. This goes as far as recognising what’s normal, unusual, or unique.
4. Bug-detector: We mean the computer and software kind, naturally. When used correctly, AI is a powerful weapon in uncovering software bugs and meeting cyber-attacks head on.

Risks when using AI in business

Regarding these “pros” - well done on being cost-effective while trying to make your business smarter and safer. But where is the margin for error and how do humans know when to check on that automated output? Here are some of the most common pitfalls and open goals that AI may produce, and some solutions on how you can protect your business from the downsides.

1)    Human error and machine-led mistakes
A simple mistake or error rather than malicious intent still takes the top spot when it comes to AI risks. A real person may initially generate this when they’re instructing an AI system on what to do, or it could come from bad output as a result of a glitch in programming and coding. But putting too much faith in an AI-based piece of software without checking for blips may well cost your business financially and reputationally, if the digital output gets derailed. Even in the fast-paced and ever-advancing world of AI, a sychstem is only as good as the foundations and information it’s built on. 

2)    Beware of the cyber-attacks
The most common, yet potentially the most destructive outcome of “negative AI”, is a cyber-attack. Whether this is in the form of a data breach, system infiltration, or bad actors managing to break through security, a cyber-attack is a very real consideration when depending on AI systems and software.

Cybercriminals have a nasty habit of knowing how to manipulate AI for their own ends and can often use their own devices to stay one step ahead with viruses and malware. This is not just some quixotic imagining where your own security systems are tilting at digital windmills. The threat is real, and however good your own AI capabilities are, it may only be a matter of time before it comes head-to-head with an opposing and dangerous AI force that could result in a release of encoded and sensitive data.

A fairly recent and worrying development falls under the insidious umbrella of social engineering where you may think you’re talking to a person, or even seeing them, but it’s fake imaging and sound replication. We’ve even seen many celebrity fakes being used to front advertising campaigns, which are often incredibly convincing but are the source of false information in the form of smoke, mirrors, and image manipulation.

As a precautionary measure, you need to be at least one step ahead of simply updating your own AI technology. To reduce the impact of cyber-criminals, it’s recommended that you take out a solid cyber insurance policy which mitigates for many of the circumstances that bad AI can lead to. 

 3) Not quite human
There are still many elements and aspects that give AI away. Chatbots are still playing catch-up when it comes to understanding and authentically replicating human responses and emotions. When we’re “talking” to “Lola from IT” on an online customer service chat, it’s not terribly difficult to figure out that Lola is in fact a pre-programmed response engine – and that can make an already frustrated customer even madder – especially when it’s an automated “voice”. After all, there’s still a human at the receiving end and many still want a breathing, sentient being to sort out what’s usually an IT issue!

We’d summarise by saying that AI is as much a part of the future as the Romans inventing proper roads, and the Industrial Revolution transforming the 19th century. We’re right to embrace the wonder of an organ transplant powered by robots and voice commands, but need to have both eyes on the open door that may let in cyber-criminals or increase the possibility of costly mistakes.

Find out more about cyber and how Howden can help your business manage both the risk and opportunity in this area visit our cyber security hub. You can also call 0330 008 1334 to talk about cyber insurance with one of our specialist team, who we can assure you are very much humans.