Cyber security is an essential part of any business. This starts with being aware of the latest threats such as AI cyber attacks. In this blog, we discuss how AI can be used in malicious ways, even by unsuspecting users. So what do you need to know about AI cyber attacks?
AI attacks occur when a malicious user asks an AI tool such as ChatGPT to create a code to steal sensitive information from a company. This can also happen unintentionally by a user asking AI tools innocent queries. 5 AI tools have already been shown to produce this type of code.
Read on to find out more about how AI can be manipulated to launch cyber attacks.
How AI Can be Malicious
There is a new upcoming threat to cybersecurity – AI generated malicious code. Vulnerabilities have been found that allow users to ask ChatGPT to produce code that can be used to hack into databases. Any sensitive information in these databases can then be stolen and sold on for the hackers profit or held hostage from the owner of the data. This malicious code can also be used to crash major computer systems and research has shown that users can accidently produce this code and cause crashes.
AI has been shown to reduce the lifecycle of data breaches by over 100 days when combined with security tools. Machine learning uses algorithms and techniques to learn from data and make autonomous decisions without having to be programmed. This means that it can learn from previous attacks and how they were resolved to find an effective solution.
However, the intelligence of AI also allows malicious users to take advantage and manipulate it into creating attacks to gain sensitive information. As well as ChatGPT, 4 other AI tools have been analysed by researchers to see if they could produce malicious codes. Each of the tools was able to produce a code that upon execution could leak confidential information to completely destroy services.
This means that AI attacks are not only accessible to expert hackers, but also the general public, even if they are not intending to be malicious. Computer systems could unintentionally be infected by the code without users even realising what their innocent actions have caused.
Common Types of Cyber Attacks
As well as the latest AI developments creating a potential threat to cybersecurity, these are other common types of cyber attacks which you should be aware of to protect your business, including:
Malware
Malware is a program or code that is intended to harm a computer, network or server. This is considered one of the most common types of cyber attack, with an estimated 72% of businesses affected by malware attacks in 2023. There are many different types of malware attack including ransomware, trojans, viruses and bots. Here are some of the common malware attacks you may have heard of:
- Ransomware – The victim’s data is encrypted and a decryption key is offered in exchange for a large payment. Sometimes, even when payment is made, the data is not retrieved.
- Trojans – Trojans are malware that leads to a legitimate looking software or free download often through bait websites.
- Adware – Spyware is used to monitor user’s activity online and produce ads that are relevant with the intention of encouraging them to click on the malicious ad.
- Botnet – This is a network of computers that are infected with malware, controlled by a ‘bot herder’ – a person who operates the infrastructure to launch attacks.
Denial of Service Attacks
Another common and potentially dangerous cyber attack is a DOS (Denial of Service) attack. This is where attacks target company networks with the intent of overloading or crashing them. Any data lost or stolen is hard to to recover as company systems can become completely unusable during the attack. The attacks work by launching an overwhelming amount of requests to the server in order for the network to crash or become unresponsive. If you are interested in finding out more about DOS attacks, check out our recent blog.
Phishing
In 2022, 82% of businesses reported that they were subject to phishing attacks, making it the most commonly reported cyber attack. Phishing attacks occur when someone with malicious intent tries to trick users into performing a dangerous action, such as clicking a link that will download malware or entering a malicious website. These actions can be completed through several different methods such as text, email and social media. To find out more about phishing attacks, read our in depth blog.
Spoofing
Criminals can pose online as a trusted source, commonly in one of the following three forms:
- Domain Spoofing – This is where attackers impersonate a trusted business or individual with a fake website or email address to appear trustworthy.
- Email Spoofing – Email spoofing is where criminals target businesses with a fake email address that appears trustworthy.
- ARP Spoofing – This form of spoofing attack is used to intercept data between a device and the intended recipient.
Identity Based Attacks
Identity based attacks are very hard to identify, meaning the time spent recovering stolen data is longer than most other cyber attacks. This is due to the attacker appearing as close to the legitimate user’s normal online behaviour. Examples of this attack are man-in-the-middle attacks, golden ticket attacks and credential harvesting.
Code Injection Attacks
This is where attackers inject malicious code into vulnerable devices and networks to change its course of action. With the new AI threats, it is expected that this type of attack will rise in frequency over the next few years as malicious code can be produced a lot faster.
Supply Chain Attacks
This type of cyber attack targets supply chains through trusted third-party vendors who offer services or software. Attacks inject harmful code into applications to infect users of a piece of software or compromise physical components in a hardware supply chain attack.
Insider Threats
Insider threats relate to current and previous employees that have access to the company systems and any data the company owns. If they have been trained on the company processes, they could create a tailored attack that leaves the business exposed.
How to Protect Your Business From Cyber Attacks
One of the best ways to prepare for and prevent cyber crime in your business is to train all employees and have policies in place for if an attack occurs. This can be through making sure only the appropriate members of staff have access to sensitive data, strong passwords including biometrics and having security software installed on all devices throughout the company.
Ensure all devices and systems used within the company are up to date to make sure they have the latest protective measures against the ever evolving cyber threats.
Expert Cybersecurity Advice from NEBRC
At NEBRC we understand keeping your business safe from attacks is one of the most important aspects of running a company. That’s why we provide expert advice against the most common cybersecurity threats including the new AI threat. Sign up for our Free Core Membership here to find out more about how we can help you protect your business.