Val GenAI Chatbot – Your virtual learning assistant

Val is a free, secure and private option for RMIT students looking to use a text-based generative AI tool. By using Val, you save money while protecting the information you provide from being shared with third parties. 

Using Val instead of public tools, even those where you pay a fee, guarantees that no one can use the information you provide. Many generative AI tools use the data provided by you to train their tool or share it with third parties. Your data is valuable and may include your intellectual property. Additionally, if you use other AI tools for assessment work there is a risk it could be sold and used for misconduct, negatively impacting you.

Val is powered by OpenAI’s GPT-4 large language model, so it has all the same functionality as GPT-4. Val has the added bonus of being private and secure, your data won’t be shared with OpenAI or other third parties. And Val is free for RMIT students to use.

Val can help you with many tasks as part of your study and personal goals.

• Use Val to give you advice on improving your writing (based on text you provide) or public speaking skills (based on situation).
• Use Val to help you explore ideas related to a particular topic or create a basic structure for an assignment.
• Paste text that contains content you are trying to learn and ask Val to quiz you on it.
• Use Val to help you plan and edit cover letters or responses to selection criteria when applying for jobs.
• Ask Val to help you draft a social media campaign, including hashtags, then use Adobe Creative Cloud software (available free to all students via RMIT Adobe Creative Campus to make it a reality!
• Need to learn to program? Ask Val to write some code for you, or provide feedback on code you’ve written, in Java, Python, Matlab, SQL, and many more languages.
• Tell Val what ingredients you have in your house and ask it for suggested recipes.

For detailed information on how Val can be used in your studies, see the self-paced Generative AI for students at RMIT module. It's packed with tips and examples to help you use AI effectively.

To access Val:

1. Click the ‘Launch Val’ button at the top of this page.
2. To start a conversation, simply interact with Val in the same way you would with any other chat application: Add a query such as a question, wait for Val to respond and then continue the conversation in a back-and-forth style.
3. To close the conversation, just close the browser tab.
4. To start a new conversation, press the ‘Start a new chat’ button (plus icon) or come back to this page and click the ‘Launch Val’ button again.

When writing a prompt, be specific and provide appropriate context and details. You can do this either in a single message or as a series of messages.

A well-structured prompt will more likely result in a more suitable result from Val. If Val does not provide a suitable response, continue to refine your prompt by providing additional information, such as definitions, the purpose or goal to be achieved, the role you want Val to take, any constraints or limitations, the format of the response or even examples.
 

Example tasks and prompts

For more tips and examples, see the self-paced Generative AI for students at RMIT module.

Follow the guidance provided in your course guide and any specific assessment instructions to know whether Val or other generative AI tools are appropriate to use in an assessment. 

Your teachers may advise you that:

• no generative AI tools can be used (this includes Val)
• a specific generative AI tool or set of tools can be used, such as Val or Adobe Firefly, because of their functionality or security
• using generative AI tools like Val is only permitted for specific parts or stages of the assessment task
• any generative AI tool can be used to support assessment completion, as long as it is used in accordance with RMIT policies and academic integrity requirements. 

You should always be transparent with your teacher regarding if, when and how you have used generative AI, and provide appropriate referencing and acknowledgement as per the RMIT Library AI referencing guidelines.

Transparency, trustworthiness, and honesty are fundamental principles of academic integrity at RMIT, and around the world.

It depends. Use of any tool in your research, including generative AI tools such as Val, must comply fully with the principles of research integrity, set out in Australian Code for the Responsible Conduct of Research. That means you are responsible for conducting your research with honesty, rigour, and transparency. 

Some uses of Val for supporting your research, such as assistance with writing code or support with representing technical text for a non-technical readers may be acceptable. Other uses for conducting research, such as asking Val to write papers or analyse data, are generally unacceptable.  

If in doubt, it is best to speak to your supervisor or a research integrity advisor. Information, resources, training, and advice on Research Integrity and connections to the Research Integrity Advisers (RIAs) in your School are all available on the RMIT Researcher Portal (accessible to supervisors and research candidates). 

For more information, please see the following resources and supports:

• A range of information and links about research integrity at RMIT can be found on the RMIT Research Integrity web page. 
• The RMIT Researcher Capability Development (RCD) team maintains extensive resources for researchers, including several online training modules for RITE (research integrity training and education). 
• An RMIT expert group has published a White Paper on Research Integrity and Generative AI.
• Research @ RMIT Vietnam maintains its own Research Ethics and Integrity information for researchers in RMIT Vietnam.

If you are a Higher Degree by Research candidate, you must declare, attribute and acknowledge use of AI in your thesis submission in accordance with the HDR Submission and Examination Procedure.

No, you should employ critical thinking to evaluate all the outputs that you receive from Val. All generative AI chatbots, including Val, may provide incorrect, biased and inappropriate content. As such, it is important to validate the information that you receive for factual accuracy against additional sources. 

Val is powered by GPT-4, a large language model pre-trained by OpenAI – it has been trained on publicly available internet data only, reflecting the predominant biases, prejudices, and misconceptions of the web’s human authors.

You should not rely on Val’s responses, especially for medical, financial, legal or any other professional advice. 

If you encounter a hateful or inappropriate response while using Val, you can report it by selecting the ‘Feedback' button at the top right corner of the Val chat interface.

RMIT’s academic integrity page provides further information about appropriate use of generative AI, such as Val.

No, Val can only access internet data up to September 2021. You should always validate the information that you receive from Val for factual accuracy against additional sources.

This is a known behaviour that occurs on versions of GPT-4, including Val. GPT-4 is trained on internet data up to September 2021, which is before GPT-4 was created. Interestingly, this means that GPT-4 doesn’t know that GPT-4 exists!

This is an excellent example of the limitations of generative AI and the importance of employing critical thinking to evaluate all the outputs that you receive from generative AI tools, including Val.

No, Val does not have access to Canvas or any RMIT Library databases.

Spell check is already a feature in many writing tools, like Microsoft Word, so it is often not necessary to use Val to spell check your writing. But Val can provide additional support by helping you with suggestions on grammar and tone, taking on the role of an editorial assistant and giving you suggestions on how to improve your writing. 

Please note, while Val is configured to use Australian English, it may sometimes provide responses in US English.

No. If asked to generate references, Val may provide something that looks like a reference or citation. But Val is not able to reliably generate true references or citations, or relate the contents of references to ideas in your work.

You should follow the RMIT Library referencing guides and tools that are intended to format references and citations in accordance with the style you require.

For detailed information on how to cite and reference content generated by Val or other generative AI tools, refer to the RMIT Artificial Intelligence Referencing Guidelines.

RMIT is collecting information from interactions with Val, including inputs (“prompts”) and Val’s responses.

This information is being collected to:

• understand how Val is being used by students and staff
• ensure Val is operating as expected and responding appropriately to prompts
• inform new features and enhancements to satisfy student and staff use.

Recorded information is de-identified and does not include student or staff IDs.

Prompt histories are stored securely for a period of six months (unless you choose to delete them) and then automatically deleted.

Information is stored in Australia, specifically in the Azure Australia East region.

For information on who can access and see information collected from Val, see the next FAQ.

The only people who can see user’s inputs and Val’s responses are a small number of authorised RMIT staff members. Educators (teachers, lecturers, etc) and other students cannot see any inputs or responses.

Recorded information is de-identified and does not include student or staff IDs. It is not shared with any external organisations outside RMIT.

Do not enter prompts that contain:

• Information that you would not usually share with other people
• Personal, sensitive and health information about yourself or anyone else. This includes details such as names, student and staff ID numbers, email addresses, phone numbers, dates of birth and photos.

No, Val does not use or learn from the data you provide. Your data will not influence the answers provided to other users. When writing prompts for Val, you may paste text and ask for an answer or task to be completed in relation to it. This does not train Val, it just provides Val with context in your individual chat to help respond to your prompts.

Use of Val must comply with the RMIT Acceptable Use Standard – Information Technology.

Student use of Val is governed by the Student Conduct Policy and Academic Integrity Policy. Use of Val by researchers is governed by the Research Policy.

Any personal information collected by the service will be handled in accordance with the RMIT Privacy Statement.

If you have any questions or concerns about how the information collected by Val will be handled, please contact aiquestions@rmit.edu.au.

If you experience an error, please try at least one of the following:

• Wait a minute or two for Val to catch their breath!
• Break a single long prompt into several smaller prompts.
• Clear the existing chat, by either refreshing the page or selecting the sweep icon next to the chat box.

If the issue persists, please report it via IT Connect.

If you experience technical issues accessing or using Val, please contact IT Connect for assistance or to report the problem.

Val is an acronym for ‘Virtual Assistant for Learning’, a nod to its potential to help RMIT students with learning and other daily tasks.