Generative AI is a type of artificial intelligence (AI) that can create new things, like stories, pictures, or music that normally require human intelligence. It does this by analyzing a huge amount of existing data (like books or images on the internet) and then mimicking and combining what it learned to create something original.

A large language model (LLM) is an AI system that uses deep learning to understand, generate, and process human language. LLMs power the generative AI tools we know today like ChatGPT. Think of them as the super brains behind a lot of the chat interfaces you see today. LLMs are trained on humongous amounts of data (like a lot of the internet) and often adhere to specific rules to protect user privacy and various laws. 

A deepfake is audio, video, or images that have been changed by AI. They are generally made using existing content (like a profile pic or YouTube video) but changed by swapping faces, cutting voices, altering images, and often make people say or do things they never did.   

AI browsers, not to be confused with AI search, are the platform that you use to access and browse the internet. These new browsers use AI to interact with you, providing customized suggestions and information versus a list of websites. AI search scans websites and provides a summary from multiple information sources. 

Cybercriminals are using AI for targeting and scale. They use AI to take all of the information about you already leaked (name, email, address and more from past data breaches) to craft individualized emails or text messages, at volume. They’re also using it to create deepfakes or voice scams. Malwarebytes research found one in five mobile users have been the target of a deepfake scam and nearly as many have encountered a virtual kidnapping scam (a decades-old tactic that now often uses AI voice cloning).  

Our detection technologies have always been powerful at identifying and blocking threats. But many of today’s most dangerous threats operate at a different level — manipulative language, deceptive intent, coordinated behavior patterns, and visual deception. AI gives us the ability to understand that context, opening up protection in areas that were previously difficult to reach.  

AI accelerates our powerful detection technologies, responding to the scale of threats and the near-perfect scams and phishing lures that AI in the hands of cybercriminals create. It fuels behavioral heuristics, spotting and blocking malicious content. It helps spot scam patterns, blocks threats in real-time and anticipates future tactics. We lead the pack at detecting and blocking evasive malware because of our AI-powered engines. 

Historically, malware research relied on human oversight to research and process, but as the volume of threats has grown, it’s become a Sisyphean task. it requires significantly more people. As cybercriminals leverage AI to create, the scale has become infinite. The need to use AI to help with the processing of threats and the creation of detections is not just a desire; it’s a need. Malwarebytes is using AI across the business to drive scale, boost productivity, enhance protections, and demystify cybersecurity for users, but we do this carefully with humans in the loop every step of the way.

We’re keeping our eye on tomorrow and continue to innovate to stay ahead of cybercriminals using AI in new and inventive ways. 

AI can provide scale, supporting important work in cybersecurity. However, we think it is very important that humans are still in the loop to make sure AI isn’t being taken advantage of by cybercriminals.