---
title: "Cheap AI Hacking: What Can Your Business Still Block?"
description: Low-cost AI agents are advancing in hacking tests. Learn what the results mean for business risk, EDR, and prevention through Isolation and Containment.
image: https://prevent-ransomware.com/hubfs/AI-Generated%20Media/Images/Cybersecurity%20Team%20Collaborating%20At%20Interactive%20Dashboard.png
---

[![CHIPS Cyber Defense Solutions, LLC](https://prevent-ransomware.com/hubfs/chips_blue_logo_higher_res-20210817212617.png) Cyber Defense Solutions, LLC](https://prevent-ransomware.com)

☰

[About](https://prevent-ransomware.com/about) [Solutions](https://prevent-ransomware.com/solutions) [Industries](https://prevent-ransomware.com/industries) [Resources](https://prevent-ransomware.com/resources) [Partner With Us](https://prevent-ransomware.com/meetings/tony-chiappetta/30-minute-initial-meeting)

[About](https://prevent-ransomware.com/about) [Solutions](https://prevent-ransomware.com/solutions) [Industries](https://prevent-ransomware.com/industries) [Resources](https://prevent-ransomware.com/resources)

[Partner With Us](https://prevent-ransomware.com/meetings/tony-chiappetta/30-minute-initial-meeting)

# Cheap AI Hacking: What Can Your Business Still Block?

![Tony Chiappetta](https://prevent-ransomware.com/hubfs/Tony%20LinkedIn.jpg)

 by [Tony Chiappetta](https://prevent-ransomware.com/blog/author/tony-chiappetta)   
November 18, 2026

What happens to your security strategy when an attacker can afford to keep trying?

Vals AI’s CyberOffense research shows an open-weight DeepSeek model completing substantial portions of simulated network intrusions. For business leaders, the concern is what affordable automation could do with the access and permissions already available inside their organizations.

**Key Takeaway:** Lower-cost attack automation could make repeated intrusion attempts more economical. Businesses should combine detection and response with controls that restrict unauthorized execution, application behavior, and access before those actions cause damage.

## So what exactly happened?

Vals tested AI agents against simulated networks. Scores measure completed objectives, not businesses’ breach probabilities.

The [original Vals CyberOffense research](https://www.vals-ai.com/benchmarks/cyber_offense) lists DeepSeek V4.1 Flash 1017 at 16.3 of 21 objectives on Nova and 21 of 31 on Lighthouse. OpenAI’s restricted-access GPT-6 Sol scored 15.7 and 27, respectively. Giving those two environments equal weight produces approximately 73% versus 81%, explaining the headline’s eight-percentage-point gap.

As reviewed November 15, 2026, Vals now includes a cloud environment, with provisional cloud scores. Three-range averages are 54.8% for DeepSeek and 65.9% for GPT-6 Sol, a different comparison.

These percentages describe benchmark objectives, not successful attacks against that percentage of businesses.

## Why does cheaper AI hacking matter to businesses?

Lower operating costs could let attackers spend more time testing weaknesses and pursuing unsuccessful paths. The business concern is the economics of persistence, not simply which model ranks first.

Vals reports roughly $2 per DeepSeek run on the first network versus about $22 for the OpenAI model. Those figures cover model API usage, not the full cost of an attack.

**CHIPS analysis:** If repeated attempts become less expensive, organizations should reconsider how much protection comes from assuming an attacker will lose interest. A neglected system, an overprivileged account, or an unnecessarily connected server can remain useful long after the first attempt fails.

For a business, a successful intrusion can mean interrupted order processing, inaccessible files, recovery labor, and diverted management attention. For an MSP, simultaneous client incidents could consume the technician capacity needed to keep other customers operating.

The research does not establish how frequently those outcomes will occur. It gives leaders a reason to examine the conditions that could allow them.

## Could EDR have detected this?

EDR may detect or block parts of an intrusion, including suspicious execution and credential-access behavior. This benchmark does not establish that a particular EDR product would succeed or fail in your environment.

Modern EDR can provide automated prevention as well as alerts, investigation, and response. Its value should be assessed through testing, coverage, configuration, and operational follow-through.

The practical question is what happens when an activity is missed, an endpoint loses coverage, or an alert does not receive attention quickly enough.

Detection and response should remain part of the security strategy. Leaders should also ask which harmful actions are prohibited regardless of whether anyone recognizes the attack.

## Can an attack be stopped without identifying it?

Yes, when an enforced control blocks an action the attack requires. The control can decide whether an action is permitted without first assigning the activity a malware name.

**What if you did not have to detect the attack in order to stop it?**

On a Windows endpoint, Isolation and Containment place boundaries around what applications may launch, access, or change. Depending on the control and policy, those boundaries can restrict unauthorized programs, constrain trusted applications, and limit access to memory, protected files, and system resources.

Think of an application as an employee with a defined job. Permission to open a document should not automatically grant permission to change protected system files or access another application’s memory.

These restrictions reduce usable Windows endpoint attack surface: the execution paths and resources an attacker can actually exploit. Blocking a required action can interrupt an attack before damaging execution, including unauthorized encryption where applicable.

Endpoint restrictions complement network segmentation and identity controls. They do not replace either.

## What changes when the attack is unknown or AI-driven?

An attacker may change the instructions, code, or delivery method while still needing familiar endpoint actions. Unknown does not automatically mean unstoppable.

A new approach may still require launching a process, abusing a trusted application, accessing memory, creating persistence, or modifying files. Changing the attack does not necessarily change the endpoint actions the attacker ultimately needs in order to succeed.

AppGuard is a proven endpoint protection solution with more than a decade of production history focused on prevention through Isolation and Containment. As explained in [CHIPS’ AppGuard overview](https://prevent-ransomware.com/AppGuard), it restricts unauthorized endpoint behavior without requiring the attack to first be identified as malicious.

That explains the prevention principle; it is not evidence that AppGuard was tested in the Vals evaluation. Protection depends on the environment, enforced policies, and attack path, and no technology stops every threat.

## What Should Businesses Do Next?

Start by testing what an attacker can actually do after gaining a foothold. Evaluate prevention and recovery alongside detection coverage.

- **Review execution freedom.** Identify where users and applications can run unnecessary programs or scripts.
- **Constrain trusted tools.** Examine administrative utilities, remote-management access, and privileged service accounts.
- **Add prevention layers.** Evaluate application control and Isolation and Containment against realistic Windows workflows.
- **Limit movement.** Separate critical systems and review third-party access and shared administrative credentials.
- **Validate resilience.** Test protected backups, restoration times, endpoint coverage failures, and incident response responsibilities.

These priorities align with the layered approach in the [CISA, FBI, NSA, and MS-ISAC StopRansomware Guide](https://www.ic3.gov/CSA/2023/230523.pdf), which includes application allowlisting, EDR, segmentation, and recovery preparation.

The objective is not to predict every attack. It is to restrict the actions an attacker needs in order to succeed.

For a practical next step, explore the [CHIPS AppGuard Video Library](https://prevent-ransomware.com/appguard-video-library) and use the demonstrations to frame questions for your own environment: Which actions are blocked, which remain possible, and what business systems could they reach?

###### Tags:

[AppGuard,](https://prevent-ransomware.com/blog/tag/appguard) [0-day,](https://prevent-ransomware.com/blog/tag/0-day) [Ransomware,](https://prevent-ransomware.com/blog/tag/ransomware) [AI](https://prevent-ransomware.com/blog/tag/ai)

![Tony Chiappetta](https://prevent-ransomware.com/hubfs/Tony%20LinkedIn.jpg)

Post by [Tony Chiappetta](https://prevent-ransomware.com/blog/author/tony-chiappetta)   
 November 18, 2026

[![CHIPS Cyber Defense Solutions, LLC](https://prevent-ransomware.com/hubfs/chips_blue_logo_higher_res-20210817212617.png)](https://prevent-ransomware.com)

AppGuard Commercial Distributor for the Americas.  
Mt. Juliet, Tennessee.

[Follow us on LinkedIn](https://www.linkedin.com/company/chips-cyber-defense-solutions-llc)

#### The Stack

- [AppGuard](https://prevent-ransomware.com/AppGuard)
- [Zimperium](https://prevent-ransomware.com/Zimperium)
- [CyberCloak](https://prevent-ransomware.com/CyberCloak)

#### Company

- [About Us](https://prevent-ransomware.com/about)
- [The MSP 3.0 Story](https://prevent-ransomware.com/MSP3)
- [Become a Partner](https://prevent-ransomware.com/meetings/tony-chiappetta/30-minute-initial-meeting)

© 2026 CHIPS Cyber Defense Solutions, LLC. All rights reserved.

Built for the Best.

```json
{
  "@context" : "http://schema.org/",
  "@type" : "BlogPosting",
  "author" : {
    "@type" : "Person",
    "name" : "Tony Chiappetta",
    "url" : "https://prevent-ransomware.com/blog/author/tony-chiappetta"
  },
  "dateModified" : "2026-11-18T10:00:00 AM",
  "datePublished" : "2026-11-18 10:00:00",
  "description" : "Low-cost AI agents are advancing in hacking tests. Learn what the results mean for business risk, EDR, and prevention through Isolation and Containment.",
  "headline" : "Cheap AI Hacking: What Can Your Business Still Block?",
  "image" : {
    "@type" : "ImageObject",
    "url" : "https://20916912.fs1.hubspotusercontent-na1.net/hubfs/20916912/AI-Generated%20Media/Images/Cybersecurity%20Team%20Collaborating%20At%20Interactive%20Dashboard.png"
  },
  "mainEntityOfPage" : {
    "@id" : "https://prevent-ransomware.com/blog/cheap-ai-hacking-what-can-your-business-still-block",
    "@type" : "WebPage"
  },
  "name" : "Cheap AI Hacking: What Can Your Business Still Block?",
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject",
      "url" : "https://prevent-ransomware.com/hubfs/chips_blue_logo_higher_res-20210817212617.png"
    },
    "name" : "CHIPS Cyber Defense Solutions, LLC"
  }
}
```

```json
{
  "@context" : "https://schema.org",
  "@type" : "BlogPosting",
  "author" : {
    "@type" : "Person",
    "name" : "Tony Chiappetta",
    "url" : "https://prevent-ransomware.com/blog/author/tony-chiappetta"
  },
  "dateModified" : "2026-11-18T10:00:00.371Z",
  "datePublished" : "2026-11-18T10:00:00.000Z",
  "headline" : "Cheap AI Hacking: What Can Your Business Still Block?",
  "image" : [ "https://prevent-ransomware.com/hubfs/AI-Generated%20Media/Images/Cybersecurity%20Team%20Collaborating%20At%20Interactive%20Dashboard.png" ],
  "mainEntityOfPage" : {
    "@id" : "https://prevent-ransomware.com/blog/cheap-ai-hacking-what-can-your-business-still-block",
    "@type" : "WebPage"
  },
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject",
      "url" : "https://prevent-ransomware.com/hubfs/CHIPS%20&amp%3B%20AppGuard%20logos.png"
    },
    "name" : "CHIPS Cyber Defense Solutions, LLC"
  }
}
```