{"id":1226,"date":"2024-06-03T13:32:37","date_gmt":"2024-06-03T12:32:37","guid":{"rendered":"https:\/\/blogs.bath.ac.uk\/academic-and-employability-skills\/?p=1226"},"modified":"2026-02-18T15:28:51","modified_gmt":"2026-02-18T15:28:51","slug":"bias-proof-your-genai-strategies-to-mitigate-algorithmic-and-human-biases","status":"publish","type":"post","link":"https:\/\/blogs.bath.ac.uk\/academic-and-employability-skills\/2024\/06\/03\/bias-proof-your-genai-strategies-to-mitigate-algorithmic-and-human-biases\/","title":{"rendered":"Navigating the GenAI Landscape: From Bias to Prompts: Key Insights and Strategies for Student"},"content":{"rendered":"

As the University embraces the power of Generative Artificial Intelligence (GenAI), we've taken proactive steps to equip you with the knowledge and skills to navigate this rapidly evolving landscape. Over the last month, using the University of Bath Instagram<\/a><\/span> account we've conducted a series of Instagram Quiz and Poll Stories, exploring various aspects of GenAI, from bias to ethics and information literacy.
\nOur initial Story on bias reached over 2,500 registered accounts, with 755 engagements. Subsequent Stories on critical thinking, data privacy, ethics, information literacy, and prompts saw even higher engagement, reaching up to 2,912 accounts and garnering up to 1,854 interactions.<\/p>\n

The quiz results revealed some insightful findings:<\/p>\n

    \n
  1. Bias Mitigation: 52% of respondents correctly identified \"asking for counter arguments\" as an effective approach to mitigate bias in GenAI.<\/li>\n
  2. Data Provenance: 64% understood that data provenance is crucial for evaluating sources and potential biases.<\/li>\n
  3. Critical Evaluation: 70% recognised the need to interrogate GenAI outputs for flaws or inconsistencies.<\/li>\n
  4. Data Privacy: 75% identified the need to corroborate GenAI output references using other reliable sources.<\/li>\n
  5. Ethics: 75% recognised that GenAI could pose risks of biased, inaccurate or harmful outputs.<\/li>\n
  6. Information Literacy: 89% identified critically evaluating outputs as the most crucial GenAI information literacy skill.<\/li>\n
  7. Prompts: 55% recognised that providing multiple examples and a clear pattern is an effective \"few-shot\" prompting approach.<\/li>\n<\/ol>\n

    However, the data also highlighted areas where students could benefit from further skills development. For instance, 29% of participants thought \"repeating the prompt to check reliability\" helps mitigate bias in GenAI tools. Additionally, 33% believed that \"providing simple, clear instructions\" would be most effective in improving GenAI output quality and validity.<\/p>\n

    Interestingly, our polls revealed:<\/p>\n