Generative AI, or GenAI for short, refers to a category of AI capable of analyzing and emulating datasets to generate various forms of content like text, images, music, videos, and code in response to given inputs or instructions. Georgia Tech encourages the thoughtful use of Gen AI technologies such as Chat GPT and Bing Chat - emphasizing the need for mindful engagement with these tools. Key aspects to consider while leveraging Gen AI tools include safeguarding information security and data privacy, adhering to legal and regulatory standards, respecting copyright laws, and maintaining academic honesty and integrity.
Some of the most common uses for Gen AI include:
Content Creation: This includes generating text for articles, blogs, and reports, creating realistic images and graphics, and even developing entire website layouts.
Media Generation: Gen AI can compose music, create new video game environments, or even generate movie scripts.
Chatbots: AI can provide personalized responses to queries, or help to schedule appointments, draft emails, etc.
Software Development: Gen AI assists in coding, by suggesting code snippets, debugging, or even writing simple programs.
Educational Tools: It's used to create educational content, like interactive learning modules, or to personalize learning experiences based on student performance.
Current Guidelines for Use of Generative AI
Encouraging the Safe Use of Generative AI
Adhere to the Academic Policy
Examine the handbooks and policies of your respective College and School, particularly those pertaining to students and faculty. It's anticipated that as the understanding of Generative AI tools' impacts grows, Schools will revise and formulate new policies. Faculty members should, in the interim, communicate explicitly to their students about the acceptable use of Generative AI in classes and academic projects. Students should also proactively seek clarification from their instructors regarding these policies as required.
Remain vigilant against phishing.
The advent of Generative AI has simplified the process for malevolent individuals to craft advanced phishing emails and produce "deepfakes" – digital creations that realistically imitate a person's voice or appearance without their permission, and on a much larger scale. It remains crucial to adhere to security best practices and to report any dubious messages to phishing@gatech.edu.
Review Content Before Publishing
Content created by AI can sometimes be incorrect, deceptive, or entirely made up (often referred to as “hallucinations”), and it might also include material that is protected by copyright. When you publish anything that incorporates material generated by AI, you bear responsibility for that content.
Protect Confidential Data
You should not input data categorized as PII (personally identifiable information, encompassing non-public research data, financial, HR, student records, medical information, etc.) into publicly accessible Generative AI tools, following Georgia Tech's Information Security Policy. Information disclosed to Generative AI tools under standard settings is not confidential and may risk exposing proprietary or sensitive information to unapproved entities.
Data that is considered Level 2 or higher in confidentiality should only be used with Generative AI tools that have undergone evaluation and received approval from Georgia Tech's Information Security and Data Privacy office for such utilization. More guidance is shared at Generative AI Draft Guidance for Privacy and Security.
Connect with OIT before purchasing any generative AI tools
As OIT explores introducing its own Gen AI service offerings, the office is actively working to ensure that any tools acquired for use by those across Georgia Tech are equipped with suitable privacy and security measures, and serve as good uses of Institute resources.
If you have already obtained, or are contemplating the acquisition of generative AI tools, or if you have any questions, please reach out to OIT's AI team at oit@gatech.edu. It is imperative for vendor-supplied generative AI tools to undergo a risk evaluation by OIT before they are put to use.