Generative AI Tools

ITSC helps HKUST members use generative AI by offering a set of tools that they can access.  Please refer to the webpage below for a list of tools available to HKUST members.

Go to Generative AI Tools

Issues of Generative AI

It's important to note that while generative AI tools offer significant benefits, they also come with limitations and require careful consideration of ethical and security implications. Human oversight and validation remain crucial to ensure the accuracy and appropriateness of the generated content.  

Hallucination  

The phenomenon of “hallucination” in the context of generative AI tools refers to the generation of information or content that is not based on accurate data or real events. This can occur when AI models, trained on vast datasets, make connections or predictions that seem plausible but are false or misleading.  As such, it’s crucial for users to be aware of this limitation and approach the output of generative AI tools with a critical eye, verifying information against trusted sources when accuracy is essential.

Data Privacy Issues

Most of the commercial generative AI tools have stated that users’ submitted data may be used for training and product enhancement.  Users are advised not to use confidential HKUST documents in those tools unless they are desensitized.  Users are also recommended to use the HKUST ChatGPT Platform which is a secure platform and your data is protected.

Legal and Copyright Issues

The legal and copyright issues surrounding the use of Generative AI tools are complex and multifaceted. Here are some key points to consider:

• Copyright Infringement: How Generative AI uses and creates copyrighted content is a contested issue. This matters especially when AI tools make new content that may look like the data they learned from.
• Ownership of AI-Generated Works: Another legal issue is who owns the works that AI produces. Normal intellectual property laws only cover works made by humans, so it’s not clear how these laws relate to content made by AI.
• Use of Unlicensed Content: Generative AI often requires large datasets for training. If these datasets include unlicensed content, it could pose legal risks for the developers and users of these tools.
• Liability: If Generative AI tools create content that is defamatory, violates rights, or causes harm, there may be questions about who is liable—the creator of the tool, the user, or the AI itself.