Natural Language Processing (NLP) For Ethical Artificial Intelligence (AI)
Bibliographic record
Abstract
The goal of my thesis is to study people’s reactions towards the ethical use of artificial intelligence. To achieve this goal, I use Latent Dirichlet Allocation (LDA) to reveal underlying topics from text data (text corpus) scrapped from AI-related subreddits. Then, on the fine-tuned topics, sentiment analysis is performed to get insight into how people react to the ethical use of artificial intelligence. The fast growth, advancements, and integration of Artificial Intelligence (AI) into the various aspects of human life have triggered widespread discussions regarding its ethical implications. Understanding public sentiment towards the ethical use of AI is quite crucial and paramount for policymakers, developers, and researchers who are aiming to ensure the responsible deployment of artificial intelligence. This thesis investigates people’s perceptions and reactions to the ethical use of artificial intelligence by analysing large-scale discussions from online communities, specifically AI-related subreddits. These platforms serve as rich sources of unfiltered public opinion, offering valuable insights into societal attitudes and concerns about this massive shift in technology. To systematically explore the complex and diverse discourse surrounding AI ethics, this study employs Latent Dirichlet Allocation (LDA), an unsupervised topic modelling technique, to uncover hidden thematic structures within the collected textual data. Now, LDA enables the identification of recurring topics (prominent topics), themes, and points of discussion by analysing word co-occurrence patterns across thousands of user-generated posts. The extracted topics are then carefully studied and evaluated, and fine-tuned to ensure topic coherence, relevance, and meaningful categorization, providing a structured overview of the key ethical issues being discussed within these online communities. The study then further conducts sentiment analysis on these fine-tuned topics to assess the emotional undertones associated with each identified topic. By quantifying sentiment polarity (positive, negative, or neutral) and emotional intensity, the research captures the nuanced reactions of the public towards different ethical dimensions of AI, such as fairness, privacy, accountability, bias, and the impact of AI on employment and human rights. This combined methodology of topic modelling and sentiment analysis offers a comprehensive framework to map not just the breadth of ethical concerns but also the depth of emotional responses surrounding AI ethics. The findings of this thesis provide empirical evidence of public opinion trends, highlighting areas where AI development is met with optimism, scepticism, or ethical alarm. By shedding more light on the topics that resonate most with the public and the sentiments they invoke, this research contributes to the broader comprehension of societal expectations and apprehensions about artificial intelligence. The insights gained can inform the development of ethically aligned AI systems and likewise help guide future public engagement strategies, regulatory policies, and educational efforts aimed at fostering a more transparent and socially responsible AI ecosystem.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.008 | 0.034 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.005 |
| Science and technology studies | 0.002 | 0.003 |
| Scholarly communication | 0.007 | 0.008 |
| Open science | 0.001 | 0.004 |
| Research integrity | 0.002 | 0.005 |
| Insufficient payload (model declined to judge) | 0.018 | 0.013 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".