Privacy Information Statement

We, a team of researchers from RMIT University (Melbourne, Australia) who created “Umibot” (also called “Umi”), take your privacy very seriously. This Privacy Information Statement governs all pages on Umibot’s website (this Website: www.umi.rmit.edu.au), and it explains what information we collect and how that information is stored, used and disclosed. By accessing this Website, you acknowledge and accept the terms of this Privacy Information Statement, Umibot’s Terms of Use and RMIT’s Privacy Statement.

The technology behind Umibot

Umi is a chatbot; that is, a computer program (not a human) that simulates conversation with users. We built Umibot, alongside our digital partner agency Tundra, using the artificial intelligence service Amazon Lex. We use Google Analytics, a web analytics service offered by Google, to track web traffic to this Website (which houses Umibot). 

This Website is hosted by Amazon.com, Inc. in Sydney, Australia.

The information we collect

Your conversation with Umibot is private. Umi’s purpose is not to collect, store, use or disclose any personal information about you. 

Umi does not record or store your conversation. No one other than you can see your conversation with Umibot.

We use Google Analytics to collect and store de-identified and aggregate information, including but not limited to:

  • Information about how Umi is used, including: the total number of users; average time and length of use; the total number and type of questions asked; the type of help sought (e.g., information about laws or support services); and the total number of times the “Quick Exit” button is used.

  • Technical information, including where exactly users left their conversation with Umibot, and bugs or other technical issues that might impact how well Umi works. 

The only information we may collect directly from you is feedback in the form of an optional and anonymous Exit Survey that is hosted by Qualtrics.

How we de-identify information

We have spent a lot of time developing our strategy for de-identifying the information that we collect about Umibot. All information is de-identified at first instance; that is, before the information is sent/passed to Amazon Lex and then Google Analytics.

For Amazon Lex, we use:

  • Amazon Comprehend that uses pattern matching to identify and redact personal or sensitive information, such as email addresses, names and phone numbers that users might disclose to us.

For Google Analytics, we use:

  • Abstract data labels that are known only to us (e.g., "Scenario 1", "Scenario 2"); and

  • Anonymise Internet Protocol (IP) addresses. 

We frequently review our de-identification strategy to ensure that it is as rigorous as possible.

What happens if personal information is shared with us

As stated above, Umi’s purpose is not to collect, store, use or disclose any personal information about you. You are not required - or encouraged - to provide your name or any information that would identify you or anyone else. However, if we receive information from you that does potentially identify you, we will keep it securely in accordance with RMIT’s Privacy Statement. 

You may seek access to any personal information we hold about you, or to update and correct that information. Please refer to the contact details below.

How we store information

We store information in hard copy documents, as electronic data or in RMIT software or systems, including cloud or other types of network storage (e.g., OneDrive) in Australia. 

We apply safeguards and take appropriate security measures to protect the information that we collect, including but not limited to: 

  • Document storage and data security processes;

  • Security measures for access to RMIT systems and networks; and

  • Website protection measures.

De-identified and aggregate information may be stored by Amazon Lex and Google Analytics on servers outside Australia (e.g., in the United States of America). However, where possible, we have chosen for data collected by Amazon Lex and Google Analytics to be stored in Australia.

When you communicate with us through a digital platform, such as Twitter or Microsoft Outlook, the platform and its partners may collect and store your personal information overseas. If we collect any personal information, we will keep it securely in accordance with RMIT’s Privacy Statement.

We will take reasonable steps to destroy the information we collect when it is no longer needed for research or other purposes.

How do we use and disclose information

We use the de-identified and aggregate information that we collect for two main purposes: 

  • Improving Umibot’s functionality to make the chatbot as helpful, safe and secure as possible; and

  • Publishing research to produce new knowledge, including concepts, methodologies and understandings for: (i) Australian and international stakeholders attempting to better detect and prevent image-based abuse; and (ii) researchers in the field - as a contribution to existing knowledge. We may disclose the de-identified and aggregate information that we collect in research outputs, such as journal articles, reports, presentations and interviews with news media and to stakeholders in Australia and overseas. You can find out more about our research at: www.imagebasedabuse.com ("the Project Website").   

We may disclose information to Australian or overseas recipients if information is unlawful or poses a threat to us or any stakeholders or in accordance with RMIT’s Privacy Statement.

Web traffic information is disclosed to Google Analytics when you visit this Website. Google may store information across multiple countries.

Cookies

We collect cookies for the purposes outlined above. Most web browsers enable users to block or otherwise manage cookies through privacy settings. In some instances, however, blocking cookies may result in Umibot not operating as optimally as possible.

Third party links

This Website may contain links to and from third-party websites. If you follow/click on a link to a third-party website, it is important to note that the website is likely to have its own privacy policy or policies. This Privacy Information Statement does not cover other third-party websites, and we do not accept any responsibility or liability for the privacy policies of third-party websites.

Privacy complaints

If you wish to make a privacy complaint about Umibot, you should first contact us via the details below. We will respond to your complaint within 10 business days and do our best to resolve the issue in a timely manner.

If you are dissatisfied with the outcome of the complaint, or the way in which the complaint was handled, then you can make a third-party complaint to the RMIT University Privacy Officer via privacy@rmit.edu.au or phone +61 3 9925 1161. 

Contact us

For all enquiries, including requests for this Privacy Information Statement to be provided in a different format, you can contact us via imagebasedabuse@rmit.edu.au or the Project Website: www.imagebasedabuse.com. You can also contact us at:

ARC Future Fellowship Project Team (ID: FT200100604) School of Global, Urban and Social Studies  RMIT University 124 La Trobe St, Melbourne, Victoria 3000, Australia Ph: +61 3 9925 2476

Changes to this Privacy Information Statement

We may at any time change the terms of this Privacy Information Statement without notice. We therefore encourage you to occasionally check back to see if anything in this Privacy Information Statement has changed. 

Your continued use of this Website will be on the terms of this Privacy Information Statement and the Terms of Use as amended from time to time.

Latest update: 02/02/24