You don’t need ChatGPT to generate a list of reasons why generative AI is often not that amazing. The way algorithms receive creative work often without permission, harbor unpleasant biases, and require enormous amounts of energy and water to train are serious problems.
However, putting all that aside for a moment, it’s surprising how powerful generative AI can be for prototyping potentially useful new tools.
I was able to witness this firsthand by visiting Club Sundaia generative AI hackathon that takes place one Sunday each month near the MIT campus. A few months ago, the group kindly agreed to let me participate and decided to dedicate that session to exploring tools that might be useful to journalists. The club is supported by a Cambridge nonprofit called Æthos that promotes the socially responsible use of AI.
The Sundai Club team includes students from MIT and Harvard, some professional developers and product managers, and even one person who works for the military. Each event starts with brainstorming possible projects which the group then narrows down to a final option that they actually try to build.
Notable launches from the journalism hackathon included using multimodal language models to track political posts on TikTok, automatically generate freedom of information requests and appeals, or summarize video clips of local court hearings to help with local news coverage.
In the end, the group decided to build a tool that would help journalists covering AI identify potentially interesting articles published in the Arxiva popular server for research paper preprints. My presence may have influenced you here, since I mentioned in the meeting that exploring Arxiv for interesting research was a high priority for me.
After coming up with a goal, the team’s programmers were able to create an embedded word—a mathematical representation of words and their meanings—from Arxiv AI articles using the OpenAI API. This made it possible to analyze the data to find articles relevant to a particular term and explore relationships between different areas of research.
Using another word embedded in Reddit threads, as well as a Google News search, the coders created a visualization that displays research articles alongside Reddit discussions and relevant news reports.
The resulting prototype, called AI NewshoundIt’s rough and ready, but it shows how large language models can help extract information in new and interesting ways. Here is a screenshot of the tool used to search for the term “AI agents”. The two green squares closest to the news article and Reddit groups represent research articles that could be included in an article about efforts to create AI agents.
Congratulations from the Sundai Club.