AI-supported spear phishing fools more than 50% of targets

Watch For Scams is dedicated to helping you avoid becoming a victim of fraud.

If you like this ezine, do a friend a big favor and forward this to them. If a friend forwarded this to you, and if you like what you read, please subscribe by visiting the link below:

Subscribe Here

AI-supported spear phishing fools more than 50% of targets

One of the first things everyone predicted when artificial intelligence (AI) became more commonplace was that it would assist cybercriminals in making their phishing campaigns more effective.

Now, researchers have conducted a scientific study into the effectiveness of AI supported spear phishing, and the results line up with everyone’s expectations: AI is making it easier to do crimes.

To this end the researchers developed and tested an AI-powered tool to automate spear phishing campaigns. They used AI agents based on GPT-4o and Claude 3.5 Sonnet to search the web for available information on a target and use this for highly personalized phishing messages.

With these tools, the researchers achieved a click-through rate (CTR) that marketing departments can only dream of, at 54%. The control group received arbitrary phishing emails and achieved a CTR of 12% (roughly 1 in 8 people clicked the link).

The research also showed a significant improvement of the deceptive capabilities of AI models compared to last year, where studies found that AI models needed human assistance to perform on par with human experts.

The key to the success of a phishing email is the level of personalization that can be achieved by the AI assisted method and the base for that personalization can be provided by an AI web-browsing agent that crawls publicly available information.

Based on information found online about the target, they are invited to participate in a project that aligns with their interest and presented with a link to a site where they can find more details.

Other bad news is that the researchers found that the guardrails which are supposed to stop AI models from assisting cybercriminals are not a noteworthy barrier for creating phishing mails with any of the tested models.

The old clues for telling if something was a phishing mail were:

1. It asks you to update/fill in personal information.

2. The URL on the email and the URL that displays when you hover over the link are different from one another.

3. The “From” address is an imitation of a legitimate address, especially from a known brand.

4. The formatting and design are different from what you usually receive from a brand.

5. The content is badly written and may well include typos.

6. There is a sense of urgency in the message, encouraging you to quickly perform an action.

7. The email contains an attachment you weren’t expecting.

When a phisher is using a Large Language Model (LLM) like ChatGPT, a few simple instructions are all it takes to make the email look as if it came from the intended sender. And LLMs do not make grammatical errors or put extra spaces between words (unless you ask them to).

They’re not limited to one language ether. AI can write the same mail in every desired language and make it look as if you are dealing with a native speaker. It’s also easier to create phishing emails tailored to the intended target.

All in all, the amount of work needed to create an effective phishing email has been reduced dramatically, and the number of phishing emails has gone up accordingly.

In the last year, there’s been a 1,265% increase in malicious phishing emails, and a 967% rise in credential phishing in particular.

Because of AI, it’s become much harder to recognize phishing emails, which makes things almost impossible for filtering software. According to email security provider Egress 71% of email attacks created through Ai go undetected.

So how do you recognize AI phishing emails?

Number 4 above—The formatting and design are different from what you usually receive from a brand—is helpful. Compare the email with any previous communications you have from the supposed sender. If there are inconsistencies in the tone, style, or vocabulary, this could indicate that the message is a phishing attempt.

Number 5—The content is badly written and may well include typos — AI phishing emails may still use generic greetings, such as “Dear user” or “Dear customer,” instead of addressing the recipient by name. Also, look for generic or mismatched signatures that do not align with the sender’s typical signature.

Number 7—The email contains an attachment you weren’t expecting— If you know the person who sent the email but don’t trust the content, contact the sender through an alternate communication method to verify whether they actually sent it.

Remember - always watch for scams!