Assignment Brief:
You work for a social media company that has been criticized by governments, non-governmental organizations, and individuals for allowing inappropriate, offensive, and controversial posts. You have been tasked with designing a rational AI agent to detect and remove such posts. A major concern of your company is the effect of curating and moderating posts on the retention and activity of the company’s users. The company’s main income is from users’ clicking on advertisements embedded in the social media stream. The agent has full access to all posts (including text and images).
(a) Your boss suggests that it is not possible to design a rational AI agent to solve this problem since computers cannot fully understand natural language (e.g. sarcasm). She argues that only humans can solve this task in a consistent and rational manner. Discuss the merits of your boss’s opinion. Maximum 200 words.
(b) Analysis of the social media problem.
(i) Based on your interpretation of the problem, determine a suitable performance measure for the social media agent that would be satisfactory to all stakeholders. Account for any ethical issues in formulating the performance metric. Maximum 400 words.
(ii) Provide a complete PEAS analysis based on your interpretation of the agent’s percepts, actions, and environment. Include a complete and concise analysis of the task environment including justification. Maximum 400 words.