DataTrendAI2024 — Lorraine Cruz

Data Trend AI (2024)
AI, Hackathon, Concept, Prototype

Data Trend AI employs artificial intelligence to detect patterns in news cycles, emphasizing its role as an analytical tool rather than an authoritative source. We critically assessed the inherent algorithmic and structural biases of these models, particularly their pretense of neutrality. By leveraging the pattern recognition and text analysis capabilities of advanced language learning models, our goal is to enhance the understanding of bias within media ecosystems.

Concept & Use Case
We are using AI to identify trends in news cycles, not as a source of authority. We are critical of algorithmic/ structural bias that is embedded into these models – and its adopting of a language of neutrality

Leveraging the pattern recognition and text analysis capabilities in advanced learning language models, this project aims to probe how machine learning technologies might enhance our understanding of bias in media ecosystems

Case Study 1: Probing GPT-4’s Assessment of
Media Bias for Israel-Palestine v1

Case Study2: The intercept/ gaza-media-bias/gaza-media-bias-headlines-of-all-relevant-articles-surveyed

Gathering Dataset
Gathered 20 articles that had Israel or Palestine in the
titles
Searched:“Israel-Palestine”
Sorting: ‘newest’
Time Frame: April 5th - April 7th
News Source: NYT, Al jazeera & Associated Press
60 data points - Just headlines –

Takeaways
Focusing the use of AI to carry out functional tasks rather than
synthesizing information
● Questioning the conceptual maps of these models, who’s histories are
being taken into account, who interests are at the center of these
data models
● Can we consider these models to have a complete holistic worldview of
the history of the world or just of those who have created it

Identify trends in large data sets
Adjusting the trends based on personal input
Using chatgpt to identify trends