Previous Meetings

Briefing Notes – First AI Ethics Discussion Group Meeting

• Create an interdisciplinary space to discuss ethical, social, and practical implications of AI.
• Explore diverse perspectives beyond technical considerations.
• Identify shared concerns (bias, accountability, transparency, democracy, future of work).

Key themes emerging from introductions:

a. AI, Bias, and Data Ethics

• Concern about uncritical trust in data (“data doesn't lie” mindset).
• Need to interrogate data origins, interpretation, and social context.
• Anxiety around ML systems shaping hiring, sentencing, and public policy.

b. Ethics of Belief & Misinformation

• AI shaping beliefs, sometimes incorrectly.
• Students relying on generative AI may harm their own learning and focus.
• Algorithms increasingly treated as unquestioned authorities.

c. AI in Law and Professional Practice

• Potential for AI to replace paralegals or assist lawyers.
• Difficulty incorporating emotion, nuance, contextual reasoning.
• Risks of hallucinated legal summaries or fabricated case references.

d. Democracy, Surveillance & Platform Power

• Social media algorithms shape political views and reinforce echo chambers.
• Concerns about censorship, biased filtering, and manipulation.
• Fear that algorithmic feeds narrow worldviews.

e. Cultural & Global Equity Concerns

• Models reflect dominant Western/white/Christian/male assumptions.
• Poor representation of African languages and histories.
• Risk of erasing minority cultures and knowledge systems.

f. Future of Work & Human Creativity

• Anxiety that heavy AI use undermines genuine skill development.
• Human output may converge toward “algorithmic norms.”
• Fear of accelerating job displacement and loss of purpose.

A. Generative AI in Writing & Education

Concerns raised:

• Students tempted to use AI for assignments, undermining learning.
• AI-produced content often generic or incorrect.
• Detection tools (Turnitin, GPT detectors) unreliable and potentially harmful.
• Overreliance erodes student confidence.

Potential benefits (cautiously acknowledged):

• Help with phrasing, idea testing, formatting.
• Brainstorming partner (“asking a person on the street”).

Proposed responses from educators:

• Shift focus from final product to writing process.
• Require disclosure of AI use.
• Design assignments requiring complex synthesis.
• Increase seminar-style learning (resource-intensive).

B. Reliability & Hallucination in AI Systems

• Contradictory claims about historical events (e.g., Vietnam War).
• Fabricated citations and legal cases.
• Falsehoods presented confidently as truth.

Ethical implication:
→ Encourages blind trust, undermines critical reasoning, and risks real-world harm.

C. Algorithmic Bias & Structural Discrimination

• Facial recognition misidentifying marginalized individuals.
• North American/colonial training dominance reproducing inequalities.
• Translation failures for non-Western languages.
• Historical erasure of colonized or oppressed populations.

Conclusion: AI risks amplifying racial, colonial, and patriarchal bias unless fundamentally restructured.

D. Capitalism, Power, and Dependence on AI

• Tech companies incentivized to maximize dependency.
• Data collection fuels attention economy.
• Acceleration of inequality and elite power consolidation.
• AI becoming “necessary infrastructure.”

E. Human Cognition, Memory & Autonomy

• Tool reliance reduces memory and problem-solving capacity.
• Cognitive dependence on AI for basic thinking.
• Risk of AI shaping beliefs and identities via personal assistants.

• Erosion of critical thinking
• Hallucinated information influencing decisions
• Bias and discrimination embedded in models
• Worsening inequality and cultural erasure
• Loss of student skills and independence
• Opaque algorithmic authority replacing human judgment
• Capitalist incentives overriding ethics

• Brainstorming assistance
• Formatting help
• Reducing cognitive load
• Creative idea generation
• Organizing information and summarization
• Accessibility support

These benefits remain conditional on transparency, critical evaluation, and context.

• Interest in continued meetings.
• Preference for later-week sessions.
• Proposal to:
  • Circulate an article before each meeting.
  • Encourage material sharing via Teams.
  • Continue exploring themes (AI & law, decolonizing AI, education, political algorithms, etc.).