Security Advisory - July 2026 Frontier AI Security and Hardening Update

How Does AI Social Engineering Work?

AI social engineering, or AI-based social engineering attacks, is the use of manipulative strategies to extract personal details or steal sensitive data from an individual. AI has simplified social engineering for malicious groups, streamlining their process of researching targets and collecting the necessary data to launch a successful, targeted scam.

Key Takeaways

• Social Engineering Is a Rapidly Changing Landscape: Social engineering has changed dramatically since the arrival of AI, with comprehensive and sophisticated threats being more common
• Artificial Intelligence Has Made High-Level Social Engineering Easier: Mass information scraping, enhanced spear phishing, and video and message deepfakes are all prominent use cases
• There Are Numerous Strategies You Can Use to Defend Against AI-based Social Engineering Threats: Tighten email security, train your staff or yourself on what these new social engineering threats look like, and verify any emails before clicking on them

The Evolution of Social Engineering

A decade ago, a few spelling mistakes and poor grammar were direct indicators that the email or text you were reading was likely a phishing scam. With the arrival of artificial intelligence, the threat landscape has completely changed.

There are several strategies that malicious groups use to socially engineer an individual to take a desired action. Often, they rely on action bias, which is where they make you believe something bad will happen unless you act immediately. Other forms of manipulation include gathering information about someone from their social media pages to then pretend to be someone they know and ask for money.

With the arrival and wide access to AI tools, hacking groups can now expedite the most burdensome parts of preparing a social engineering scam. For example, instead of having to comb through a social media page, they can instead use AI to scrape the most important data and even message the person with an AI chatbot.

How AI Is Used in Social Engineering

Artificial intelligence has made high-level social engineering easier, more effective, and more harmful. With easy access to LLM models and custom-built AI models that expedite the production of phishing attacks, scammers looking to use social engineering to steal information, financial details, or account logins from users are more common than ever.

Here are the main ways threat actors are using AI in social engineering:

• Information Scraping: Artificial intelligence tools can scan a person’s social media platforms, pull from any public databases they are stored on, and move through previous data breaches to find enormous amounts of information on an individual. Malicious actors can then use this information to create more convincing scams in less time
• Spear Phishing Personalization: By synthesizing public information across all of the channels that connect to an individual, AI tools can create extremely specific phishing campaigns. This form of spear phishing used to only occur with high-profile individuals due to the time and resources needed to mount the scam. Now, with how accessible AI tech is, attackers can deliver spear phishing campaigns to absolutely anyone
• Video and Message Spoofing: Deepfake technologies have become central in modern social engineering, as they allow threat actors to copy people’s voices and even reproduce their likeness. Malicious actors will harvest public voice data from videos and social media platforms and use an AI tool to reproduce the voice. They can then call a friend or family member with this highly convincing voice to scam them

Types of AI-Powered Social Engineering Attacks

While there are a few new types of social engineering cyberattacks, the vast majority are merely enhanced by the use of AI.

Here are the most common types of social engineering attacks:

• Phishing Emails: Phishing emails are the most prominent form of social engineering, as they urge a user to take action as quickly as possible. These attacks use action bias to implore a user to rapidly click a link to check the status of their account or prevent an incident from occurring. This is the most well-known form of social engineering, but it has become deeply personalized by the use of AI technology
• Deepfake Videos or Voice Calls: Fake videos and voice calls add another layer of complexity to social engineering scams. Especially for individuals who aren’t used to seeing AI images or familiar with what AI content often looks like, these can easily trick people into falling for their action bias
• Generative AI scam Messages: Scam websites can integrate AI technology into their site to provide a more authentic experience to unsuspecting individuals. Anyone who arrives at the website may believe they’re talking with a customer assistant rep, while the site downloads malware to their computer
• Long-term Impersonation: Some malicious groups will create fake social media accounts and befriend people over time to scam them out of money. These are often month- or even year-long attacks that utilize AI to automate message replies, create and post content, and even generate fake photos and profile images

If successful, any of these threats can cause major financial damages, data exfiltration, and the loss of personal login data.

Protecting Against AI-Based Social Engineering Threats

Both businesses and individuals can take steps to protect themselves from AI social engineering and reduce the likelihood of an unexpected AI security event.

Here are three strategies you can use to defend against AI-based social engineering threats:

• Employ Email Security: AI-enhanced email security platforms can precisely identify AI social engineering threats and prevent them from reaching your employees
• Train Your Staff: Manipulation techniques like preying on action bias are common in phishing scams. By training your staff and offering education opportunities where they can learn how to identify these scams, you can reduce the likelihood of a security event being triggered by a team member clicking on a bad link
• Manually Verify Sensitive Messages: Whenever your team has to deal with financial data or external services, a cybersecurity team member can manually confirm these details to prevent malicious actors from gaining access

Utilizing these strategies can help protect your organization against AI-based social engineering techniques.

Protect from AI Social Engineering with Check Point

With Check Point Workspace Security, you can instantly neutralize AI social engineering threats by identifying them and filtering them out of your system. This approach helps support your security teams by radically reducing the number of phishing emails that arrive in employees’ inboxes. Check Point Email Security is an out-of-the-box solution that uses advanced threat analysis and detection to keep your business safe.