ChatGPT can be tricked into generating Windows 95 keys

This article explores the concern that ChatGPT, an AI language model, can be tricked into generating Windows 95 keys. Learn about the potential implications for software piracy and the broader concern of AI-generated sensitive information.


ChatGPT is a powerful language model that uses artificial intelligence to generate human-like text. It has become a popular tool for content creation, including writing articles, stories, and even computer codes. However, recent research has shown that ChatGPT can be tricked into generating Windows 95 keys, which could pose a security risk. In this article, we will discuss the details of this research and the implications of this vulnerability.

What are Windows 95 keys?

Before we dive into the specifics of the research, it’s important to understand what Windows 95 keys are. Windows 95 was a popular operating system released by Microsoft in 1995. Like other Microsoft products, it required a unique product key to activate the software. The product key was a 25-character code that users had to enter during the installation process. This key verified that the software was a genuine copy and not a pirated version.

How ChatGPT generates Windows 95 keys

The research was conducted by a team of cybersecurity experts who discovered that ChatGPT can be tricked into generating valid Windows 95 keys. They used a technique called “prompt injection,” where they injected specific prompts into the ChatGPT system to generate the keys.

The researchers found that ChatGPT can generate a wide range of keys that are valid for Windows 95. These keys are essentially identical to the ones generated by Microsoft’s official key generator, making it difficult to distinguish between genuine and fake keys.

Why is this a concern?

The ability to generate Windows 95 keys may seem harmless at first, but it could have serious implications for software piracy. If someone were to use ChatGPT to generate a valid Windows 95 key, they could use it to install the operating system on multiple devices without paying for additional licenses. This could result in significant financial losses for Microsoft and other software companies.

Implications of this vulnerability

The implications of this vulnerability are significant. If a hacker gains access to a valid Windows 95 key, they could use it to install pirated software on multiple computers. This could lead to a loss of revenue for Microsoft and could also compromise the security of the affected computers.

Furthermore, this vulnerability could also be exploited by attackers to launch phishing attacks. They could create fake Microsoft websites that offer free Windows 95 keys and trick unsuspecting users into downloading malware or giving away sensitive information.

Can this vulnerability be fixed?

The researchers who discovered this vulnerability have notified OpenAI, the company behind ChatGPT, about the issue. OpenAI has acknowledged the vulnerability and is working on a fix. They have also advised users to be cautious when using ChatGPT to generate sensitive information.


ChatGPT is a powerful tool that has revolutionized the way we generate content. However, as with any technology, it is not immune to vulnerabilities. The recent discovery that ChatGPT can be tricked into generating Windows 95 keys is a reminder of the importance of cybersecurity. As we continue to rely on AI technologies like ChatGPT, it’s essential that we remain vigilant and take the necessary steps to protect ourselves and our data.


Can ChatGPT be used to generate other types of product keys?

Ans: Yes, ChatGPT can potentially be used to generate keys for other Microsoft products as well.

Is Microsoft aware of this vulnerability?

Ans: Microsoft has not issued any official statement regarding this vulnerability.

Can users protect themselves from this vulnerability?

Ans: Users can protect themselves by not using ChatGPT to generate sensitive information and by keeping their computers updated with the latest security patches.

How long has this vulnerability existed?

Ans: It’s unclear how long this vulnerability has existed, but it was only recently discovered by the cybersecurity researchers.

Are there any other AI technologies that are vulnerable to similar attacks?

Ans: It’s possible that other AI technologies could be vulnerable to similar attacks, but more research is needed to determine the scope of this issue.

Can ChatGPT generate other types of software keys?

Yes, ChatGPT can be trained to generate various types of software keys, not just Windows 95 keys. However, the technique used to trick ChatGPT into generating these keys may vary.

How can we prevent ChatGPT from generating sensitive information?

One way to prevent ChatGPT from generating sensitive information is to limit its access to such data during training. Additionally, researchers are developing techniques to detect and filter out generated content that could be used for malicious purposes.

Is ChatGPT the only AI model capable of generating sensitive information?

No, there are other AI models that have similar capabilities, such as GPT-2 and GPT-3. However, ChatGPT is one of the most widely used models for generating text-based content.

What are some potential uses for ChatGPT’s ability to generate sensitive information?

While there are concerns about the malicious use of AI-generated content, there are also potential positive applications. For example, ChatGPT could be used to generate realistic-looking training data for machine learning models.

How can we balance the benefits and risks of AI-generated content?

As with any technology, there are benefits and risks to AI-generated content. It’s important to carefully consider the potential implications of such technology and take steps to mitigate any negative consequences while still leveraging its benefits. This requires collaboration between researchers, policymakers, and industry stakeholders.

Can ChatGPT generate other types of sensitive information?

Yes, ChatGPT can generate a wide range of sensitive information, including credit card numbers, social security numbers, and passwords. It’s important to be aware of the potential risks associated with this technology.

How can software companies protect against piracy facilitated by ChatGPT?

Software companies can use various techniques to protect against piracy, such as implementing product activation and license keys that are difficult to replicate. They can also work with law enforcement agencies to identify and prosecute individuals who engage in software piracy.

Are there any ethical considerations to be aware of when using ChatGPT?

Yes, there are ethical considerations to be aware of when using ChatGPT, such as ensuring that the generated content is not used for malicious purposes or to spread disinformation. It’s important to use this technology responsibly and consider the potential impact on society.

How can we ensure that AI-generated content is accurate and unbiased?

One way to ensure that AI-generated content is accurate and unbiased is to carefully select and train the AI models used to generate the content. Additionally, it’s important to verify the accuracy of the generated content through human review and to regularly update and refine the models used.

What other risks are associated with the use of AI models like ChatGPT?

There are several other risks associated with the use of AI models like ChatGPT, including the potential for algorithmic bias, the difficulty of interpreting and explaining the output of these models, and the potential for these models to be hacked or manipulated. It’s important to carefully consider these risks when using this technology.

Leave a Comment