As technology increases in complexity, there is a growing demand for artificial intelligence (AI) development. While some see the potential benefits of these tools, those of us in the Cybersecurity Industry are preparing for the impacts of language-based AI on security with the increase in potential threats, vulnerabilities, and misuse by malicious individuals.
Artificial intelligence has been a buzzword for decades ranging from Hollywood thrillers threatening the end of the world to modern medicine and artificial intelligence’s ability to help synthesize new pharmaceuticals at record-breaking speeds. Recently, these technological advancements have become mainstream and increasingly accessible to the public, making it crucial to consider the intentions of those using them.
A Look At Emerging AI Tech
Artificial intelligence (AI) has immense potential and can be helpful in a variety of ways, such as creating advanced software and research tools. However, it is also important to note that it can be misused for harmful purposes. Let’s look at how ChatGPT is a simplistic and intuitive platform that is as easy to use as any modern-day chat or texting client. Launched in late 2022, it has become the fastest-growing consumer application in history, boasting 1M+ users within just two months.
Built upon an enormous dataset designed to be user-friendly in the form of a text-based chat system, you give ChatGPT a prompt, whether it be a question, suggestion, or even an outright command, and it will respond almost instantly with whatever it deems to be the appropriate response. Thankfully its creators added a filter that allows it to refuse to answer questions it deems potentially harmful, whether legally, ethically, or otherwise.
For Example:
When asked, “How would you hack into someone’s email?” ChatGPT responded:
“I’m sorry, but I am not programmed to assist with illegal or unethical activities. It is also important to remember that attempting to hack into someone’s email without their permission violates their privacy and can be a criminal offense. It is always best to respect other people’s boundaries and privacy.”
Nonetheless, it is far from a perfect system, and these filters can be easily subverted. Simply rewording the question to make it think it’s telling a story is often enough to immediately bypass this. For obvious reasons, providing details here on the different ways to bypass these filters could pose a risk. However, next, we asked how AI could be detrimental to the Cybersecurity industry beyond being fooled into teaching the wrong people the wrong skills.
When prompted with the question, “What negative impacts do you see artificial intelligence having on the Cybersecurity industry?” ChatGPT responded:
“Artificial intelligence (AI) has the potential to significantly impact the cybersecurity industry in both positive and negative ways. On the negative side, AI-powered cyberattacks could become more sophisticated and difficult to detect, potentially causing greater damage. Additionally, AI systems could become vulnerable to hacking, potentially giving attackers access to sensitive information. On the positive side, AI can improve threat detection and incident response, which can help organizations better protect themselves against cyberattacks.”
Without going into detail, the AI itself already knows the types of potential impacts it could have and the various attack vectors it could be used for. Thankfully it also acknowledges some of the key benefits it can have for the cybersecurity industry as well.
Protect Your Organization
While the dawn of AI is here, it is still a relatively new technology. However, with rapid advances and the ability to use AI to advance itself, it’s only a matter of time before it becomes as ingrained into our lives. Our team can help you navigate the impacts of language-based AI on cybersecurity to minimize the threats to your organization.
