AI has been the hot topic for over a year now, with new developments almost monthly. AI and cybersecurity will naturally be two intertwined topics – let’s go over how the fields intersect and what the key takeaways are.
Artificial intelligence in the IT sphere
Starting with ChatGPT – the first large-language model to become a household name – AI has begun taking over more of a role in various industries. Having started with text and image generation, as an idea creator, or a translator tool, inroads have since been made into other fields including code-generation.
This integration with more and more fields and apparent utility has gradually made AI a buzzword. Institutions looking to appear more modern will use or claim to use AI in their systems – from businesses all the way to state governments. But this can expose them to some unexpected side effects:
Cybersecurity impacts of AI
As noted in this article as well as some of our previous posts, AI can and will be used in order to automate a variety of cyber threats, making them harder to detect and more prolific. We can expect a surge in AI-driven phishing attacks that closely mimic human communication in so-called spear phishing attacks.
Besides this most direct impact, however, there’s a number of secondary or indirect impacts.
- Irresponsible data access – AI training data sets have already included classified or sensitive information, and AI tools have an unfortunate tendency to get access to information they shouldn’t access. When those tools are compromised, this data is compromised in tandem.
- Opaque data handling standards – Most LLMs are a black box in terms of data going in and data going out. DeepSeek is the poster child for this, with possible data leaks to China.
- Insufficient security – AI tools are being treated as more than mere software, and not being scrutinized for security flaws in the same way as other software.
At the same time, AI will also play a crucial role in the arsenal of defense teams. AI and ML algorithms will be employed to analyze and process vast amounts of security data, enabling the identification of anomalies and providing real-time threat predictions. Only time will tell if these factors will balance out the negatives stated above.
Your Data In Your Hands – With TECH-ARROW
