# AI Is Speeding Up Cybercrime - Threat Wire

Source: https://www.youtube.com/watch?v=ndgTuk0GVOo
Recap page: https://rapidrecap.app/video/ndgTuk0GVOo
Generated: 2025-09-03T18:02:43.049+00:00

---
## Quick Overview

The video discusses how AI is accelerating cybercrime, highlighting the discovery of "Prompt-less" AI-powered ransomware that can generate malicious code on the fly and evade detection, and also details a Salesforce data breach impacting many companies due to vulnerable integrations.

**Key Points:**
- Researchers discovered the first AI-powered ransomware, dubbed 'Prompt-less', capable of generating malicious code dynamically.
- This new type of ransomware uses AI models like LLama to evade traditional security measures.
- The ransomware can tailor its attacks, making it more effective and harder to detect.
- A separate incident involved a data breach affecting Salesforce customers due to vulnerable integrations with third-party companies like SalesLoft.
- The Salesforce breach impacts numerous companies, and authentication tokens stored or connected to the affected platforms are considered compromised.
- The video also references a previous report on AI misuse in cybercrime, including AI-generated ransomware and the sale of AI-powered malicious tools on the dark web.

![Screenshot at 01:52: A screenshot of the ESET Research article titled "First known AI-powered ransomware uncovered by ESET Research", illustrating the topic of AI-driven cybercrime.](https://ss.rapidrecap.app/screens/ndgTuk0GVOo/00-01-52.png)

**Context:** The video explores the increasing sophistication of cyber threats driven by artificial intelligence. It highlights two key incidents: the emergence of AI-powered ransomware that adapts its code in real-time and a significant data breach linked to Salesforce integrations. These events underscore the growing challenges in cybersecurity as malicious actors leverage advanced AI technologies.

## Detailed Analysis

The video from "Threat Wire" discusses the growing trend of AI being used to enhance cybercrime. It starts by detailing the discovery of a new type of ransomware, "Prompt-less", which utilizes AI models like LLama to generate malicious code dynamically, making it difficult to detect and defend against. This AI-driven approach allows the ransomware to adapt its attacks in real-time. The report emphasizes that this is the first known instance of AI-powered ransomware capable of such adaptive behavior. The video then shifts to a separate, but related, security incident: a data breach affecting Salesforce customers. This breach was caused by vulnerable integrations with third-party companies, specifically mentioning SalesLoft. The compromise means that authentication tokens stored in or connected to the SalesLoft platform are potentially compromised, impacting a wide range of companies. The report also touches on previous findings about AI being used to automate various stages of cybercrime, from phishing to writing ransomware code, and its sale on the dark web, noting that AI is enabling cybercriminals to operate more efficiently and effectively.

### AI-Powered Ransomware

- Discovery of 'Prompt-less' ransomware
- Uses AI models like LLama for dynamic code generation
- Evasion of traditional security measures
- Real-time adaptation of attacks

### Salesforce Data Breach

- Impact on Salesforce customers due to vulnerable third-party integrations
- SalesLoft mentioned as a vendor
- Compromised authentication tokens for affected users

### Broader AI Misuse in Cybercrime

- AI used for phishing, ransomware development, and reconnaissance
- AI-powered tools sold on the dark web
- Increased efficiency and speed for cybercriminals

### Research Findings

- ESET Research identifies new threats
- Google Threat Intelligence Group reports on AI misuse
- Anthropic's research on AI security and misuse

![Screenshot at 00:00: Host of the show, Sarah, introducing the topic of cybersecurity threats.](https://ss.rapidrecap.app/screens/ndgTuk0GVOo/00-00-00.png)
![Screenshot at 01:52: Screenshot of an ESET research article about AI-powered ransomware.](https://ss.rapidrecap.app/screens/ndgTuk0GVOo/00-01-52.png)
![Screenshot at 02:25: Screenshot of a report detailing the misuse of AI in cybercrime.](https://ss.rapidrecap.app/screens/ndgTuk0GVOo/00-02-25.png)
![Screenshot at 03:47: A user comment discussing the problems with the "Marque and Reprisal system" and the need for investment in better infrastructure.](https://ss.rapidrecap.app/screens/ndgTuk0GVOo/00-03-47.png)
![Screenshot at 04:47: Host referencing a "DefCon vlog" about AI and cybersecurity.](https://ss.rapidrecap.app/screens/ndgTuk0GVOo/00-04-47.png)