Your Project Ideas

Below are all of the projects we have received so far in order of submission date. Each project is attached to its theme icon, plus one symbol (square, circle, diamond) to indicate the participation mode:

🟦
Canary Wharf
only

🟡
Online
only

🔶
Mixed
Mode

GenAI, Mixed Mode

Designing a workflow to integrate Gen AI into the Qualitative Research Process

Description

There has been considerable interest in integrating the power of Gen AI into the research process, but as is often the case, the early focus has been on quantitative methods. We should take this opportunity to ensure that students and researchers who use qualitative methods are not left out in the cold and investigate how Gen AI can be integrated - end to end - in qualitative research practices.

We will look at the systematic integration of Gen AI at every stage of qualitative research, from research question formulation, through literature searching, text analysis and reporting.

I can give guidance in the use of qualitative/text analysis techniques and in using popular qualitative analysis tools.

Jim, Digital Skills Development

GenAI, Mixed Mode

AI and Ethics Game

Description

A game can be a fun way for UG students to learn about the ethics of AI and explore the various issues with using AI in academia. It needs to be fairly simple and take no longer than 30mins to play to allow time for it to be taught in class. 

MIT created and are playtesting a card game designed for school students. AI Tiny Games created a simulation on ChatGPT called ‘AI Ethics challenge: Society needs you’ but you need the paid for version of ChatGPTPlus to play. ARC created a version of the card game Fiasco but it was designed from the developer/engineers perspective and I would like to create something from a user perspective.

My game will be aimed at University Undergraduate students and will focus on the ethical issues with using AI in academia. I would like it to be used by courses to introduce students to the pros and cons of AI and start a discussion about some of the ethical issues with gen AI.

What can you gain from participating in this project?
I have been involved in projects to create games for Undergraduates before and I am I can share some different types of games that have been developed for education that use different game mechanics. I can teach people how to use various tools that might be useful to create a game such as H5P and Twine.

I can also help to collate data and statistics about AI which could be used in the game.

Geraldine, Digital Education

Assessment & Feedback, Canary Wharf only

Assessgram

Description

The problem is students do not see themselves a part of assessment process. That is why they do not improve their self assessment skills. However, self assessment is a key for academic success. We need a website where students have academic profiles. They post about their current studies and academic success. Most importantly, they r gonna assess each other's performance on weekly basis . They can compare their grades, time spent on tasks, and overall grade progress. Chat section will allow them to find friends outside their class and even school. 

In UCL, we have forums and we r required to discuss some things but it is also observed by teachers so many dont get involved. As most of us are international friends, we r hesitant becoming friends at the first semester. If we saw each other's academic profiles, we could get more info about them and feel more willing to chat. Our generation is better at online interactions.

Tbh I dont have anyyy programming skills. However, I am good at making presentations and convincing people that I am right :) so i think i can present the product. I think I am fast at collecting data needed

Gunel, studying Curriculum, Pedagogy and Assessment (CPA) at IOE

Assessment & Feedback, Canary Wharf only

Using AI for automated draft formative feedback

Description

Formative feedback is valued by students as they are able to use the feedback to identify areas that need more work. In structured assignments where an exemplar solution can be created, AI tools may be able to provide specific pointers on how the student’s submission compares to the exemplar solution using PDF comparison. My goal for this AI tool would be to highlight conceptual differences in the PDFs that would account for marks, not simply a similarity checker. An alternative may be to use a marking rubric and have an AI highlight sections of the submission that are applicable to the section. I would like to explore the idea of using AI tools to enhance the process of providing or drafting formative feedback, the practical tools available inside and outside UCL, and the ethics and perception of such tool.

Possible starting points:

There are some document parsing tools, such as Google Document AI (try it here: https://cloud.google.com/document-ai/docs/drag-and-drop) and Amazon Textract, which can extract relevant sections of text based on some search keywords. ChatPDF (https://www.chatpdf.com/) is a tool where you can ask questions about the PDF in a similar format to ChatGPT. Glide (https://www.glideapps.com/ai) provide a no-code method of developing an app with AI capabilities. These tools all provide a potential foundation to build an application to process submissions to automate drafting of formative feedback.

I have some basic technical experience with working with LLMs and can share my practical experience using ChatGPT, BardAI, Github co-pilot, and other generative AI tools. I can also get you started with developing and using AI in Python/R programming and thinking about integrating AI tools in your workflow.

Kevin, Institute of Health Informatics


Assessment & Feedback, Canary Wharf only

Let's co-design MyPortfolio's homepage

Description

MyPortfolio, is our name for the open source platform, Mahara which is UCL's e-portfolio system. The challenge is to review and re-design the homepage and to align it with other UCL educational technologies such as Moodle. We will focus on creating wireframes and mock-up ideas for how to structure the homepage. Don't worry if you haven't use the tool. You will get time to learn about it on the day.

Collaborating with users / members from a community to co-design or co-create products or process is a growing practice across many sectors such as urban planning, healthcare and education. People who have first hand, lived experience of a particular service, tool or problem can offer valuable insights into developing solutions. When it comes to assessment practices, students and staff have essential experience which can help shape the home-page of our e-portfolio tool. The 'Dementia Citizens' is an example of an impressive co-design project https://www.nesta.org.uk/project/dementia-citizens/

What can you gain from participating in this project?
As a learning designer and technologist I have worked on numerous co-design projects with staff and students. My research and learning design experience will help the group learn about practical considerations and approaches for reviewing technology with users and putting people's experience at the heart of the design process.

Irina, Digital Education

Optimisation of learning the research protocol with GenAI  🏆

Description

Reading the material and methods section of a research paper requires expertise in the field and depends on the clarity of the information. Often, crucial information for repeating the experiments is omitted leading to non-reproducible experiments and loss in research funding and resources. 

The aim of this project is to create a GenAI tool for screening similar papers of interest. Thus, helping scientists to identify papers with the most suitable and repeatable protocols. Also, by comparing the most repeatable steps of an experiment usually hidden from human eyes, GenAI can help screen and detect the steps that are most consistent in a certain experiment. This could reduce the time and resources needed in R&D and allow for reallocation of these to generation of ideas, discoveries, and innovation.

Possible starting points
There are currently no benchmarking tools, we could create LLM ( Large language models e.g., ChatGPT) plugins using or similar to the new GPT store function to extract crucial experimental information and this could be combined with a Uber-style model for estimating experiment time and resources needed to guide students. Myaidrive (https://myaidrive.com/), a tool for reading research papers with AIPDF plugins are possible candidates that can be integrated for screening. The text output can then be passed through a “Natural Language Processing” protocol to systematically extract crucial information. An example of the method has been used in https://www.protocols.io/ , a website that contains published protocols as well as project planner (management) for users.

However, it (https://www.protocols.io/) doesn't seek out or filter existing protocols organically and depends heavily on information fed into it by the users.

We will provide idea development, collaboration, and business pathways as well as research experience. Data for generating robust protocol are mostly open-access publications.

Amy and Tope, Medical Sciences