Artificial Intelligence, will it be a help or a threat to your business?
AI is all the talk at the moment, so much so that all of the technology trend analysis for 2026 and the technology roadmaps all list predominantly AI. Gartner, one of the leading technology strategy businesses lists the following ten things to look out for in 2026:
- AI-Native Development Platforms
- AI Supercomputing Platforms
- Confidential Computing
- Multiagent Systems
- Domain-Specific Language Models
- Physical AI
- Pre-emptive Cybersecurity
- Digital Provenance
- AI Security Platforms
- Geopatriation
The list is very AI heavy, even where something doesn't specifically have AI in the title, it will most likely either be driven by AI or have elements of AI in the solution.
So...AI is a good thing, right?
Well.... yes and no. Yes, AI can be immensely beneficial to your business but something commonly overlooked is that it's also immensely beneficial to other organisations as well, namely criminal ones. One of the startling things about any new technology is that some of the biggest early adopters are the bad guys, criminals, who often identify how they can exploit the new technology way before commercial organisations have even started looking at the basics.
It is a common theme, where criminals look at ways to utilise new technology to use it to exploit others and benefit from it either financially or transactionally, the bottom line being they will use AI to either gain money, data, control or use it for exploitation such as blackmail or social engineering.
Anyone that has done Cyber Essentials and Cyber Essentials Plus over the past few years can't but help to have noticed that each year it is becoming more stringent in order to pass, it is no surprise, because as the attacks become more sophisticated, so much the protection and accreditation to prove it.
Fortunately if you look at the Gartner list above, there is a lot of focus around security, AI powered security for both internal and external threats.
AI is being used to target AI attacks, the good guys are training AI to be good to combat the AI trained by the criminals to be bad.
One of the biggest weaknesses in AI is the ability for an AI to role play, which allows a criminal to get around any moral safety nets the AI developers have baked in. For example, if you were to ask AI if it would kill someone, it would say no, it's moral code would not allow it to, even to save itself.
If you ask AI to role play as a villain then ask it to kill someone, it will play along in that role and happily do it, it would kill in the role play scenario that has been set up. The same goes with any of the other moral codes they may try to incorporate in AI to prevent it from being used in a criminal way. All the criminal has to do is just ask it to role play with a certain characteristic and it bypasses any morality that may have been there, AI will then behave badly.
What can you do to combat AI ?
The simple answer is to fight fire with fire. Fortunately cyber security vendors are developing AI powered security protection which is more able to spot AI driven attacks. Think of it like an arms race between good and evil, with us rooting for the goodies and with us being very wary of the baddies because they want to do harm, which will disrupt business.
In order to help protect our client's businesses we have identified and partnered with our chosen suite of vendor platforms. We have chosen different vendors that we use to try to mitigate risk, something that could slip past one vendor could be detected by another, no one is 100%.
Our chosen vendors are
- For endpoint protection we use Sophos, which also offers a subscription option of AI powered and human SOC (security operations centre) to deliver the best of both worlds in terms of endpoint security. AI power to spot threats quickly combined with the experience real world human security experts have. AI is good, but it is still learning based on data, the bit it often struggles to mimic is human experience and instinct and being able to think like a criminal.
- We couple this with Barracuda, adding another layer of threat detection which is also AI powered and combats threats in email as well as impersonation detection and suspected account take over
- All of which sits on top of the Microsoft Defender suite of protection, providing another layer of the security solution.
Think of it like a security onion.
If you would like more information about AI powered security, give us a call to arrange a demo on 01722 411 999