๐ฃ๏ธ Week 11 - Class Roadmap (90 min)
2023/24 Winter Term
OBJECTIVE: Discuss the implications of Machine Learning models in the real world.
๐ฃ๏ธ Roadmap
๐ฐ Generative AI and misinformation (~30 min)
Generative AI algorithms are prone to hallucination. While this can be fun at times, it can also be dangerous.
In this part of the class, your task will be to create a fake news article using a generative AI tool (e.g ChatGPT, Google Bard or Bing AI Chat). You will then show us your article and we will discuss the implications of this technology.
๐ก๏ธ Generative AI ethics and security (~30 min)
In this part of the class, weโll be having look at these articles:
- George Hurlburt, โWhat If Ethics Got in the Way of Generative AI?,โ in IT Professional, vol. 25, no. 2, pp. 4-6, March-April 2023, doi: 10.1109/MITP.2023.3267140.
- Matt Burgess. โGenerative AIโs Biggest Security Flaw Is Not Easy to Fixโ. Wired UK. 06 September 2023. https://www.wired.co.uk/article/generative-ai-prompt-injection-hacking.
- Thomas Claburn. โAI Assistants Help Developers Produce Code Thatโs Insecureโ. 21 December 2022. https://www.theregister.com/2022/12/21/ai_assistants_bad_code/.
- Siladitya Ray. โSamsung Bans ChatGPT Among Employees After Sensitive Code Leakโ. Forbes. 2 May 2023. https://www.forbes.com/sites/siladityaray/2023/05/02/samsung-bans-chatgpt-and-other-chatbots-for-employees-after-sensitive-code-leak/.
Instructions:
- In this section of the class, youโll be separated in groups and made to pick an article to read and discuss between yourselves (~15 min)
- Youโll then share your insights and reflections with the rest of the class in a more general discussion.
๐ Reflection and Summative Help (~30 min)
In this final part of the class (and of this course), the class teacher will first do a brief recap on the bibliographic search techniques that were supposed to be covered in the week 10 lecture but werenโt.
The rest of this section will be dedicated to reflection and help with the summative assignment.