iiinova — where your digital self keeps evolving

Seth Bernstein

@sethbern.bsky.social

1.1K followers 873 following 11 posts

PhD Student at UMich | Researching #HCI, #ComputingEd, and #genAI (he/him/his)

Seth Bernstein @sethbern.bsky.social · 08/10/2025

What’s missing? The social side. Equity, belonging, and collaboration barely show up in the data even though they’re what determine who feels welcome in CS and who walks away.

Harms as a percentage of each population. 69.6% of instances involved introductory students, 12.0% upper-level undergraduates, 9.3% graduate students, and 9.1% educators

Percentage of harm instances by evidence type within each type of GenAI tool.

130

Seth Bernstein @sethbern.bsky.social · 08/10/2025

Only about 1 in 5 papers from 2024 even mention harms. And most discuss hypothesized harms, not ones that were observed or measured. It’s time for research that examines what’s really happening in classrooms, not just what we think might be.

The stacked bar chart shows the proportion of papers based on their highest level of evidence. Papers from 2025 are omitted from this visualization.

110

Seth Bernstein @sethbern.bsky.social · 08/10/2025

Most papers focus on benefits such as faster feedback, better explanations, higher grades. But the downsides are real too: • Shallow learning • Over-trust in wrong outputs • Less collaboration • Uneven access

Stacked bar chart showing the percentage of papers that discussed harms of GenAI in computing education. Greyed bars represent papers not analyzed. Papers from 2025 and other excluded works are omitted.

110

Seth Bernstein @sethbern.bsky.social · 08/10/2025

Everyone talks about how generative AI helps students. Almost no one asks how it might hurt them. We reviewed 224 papers on GenAI in computing education to see what’s being ignored. Here’s what we found at the @templehci.bsky.social👇 📄 “Beyond the Benefits” — Koli Calling 2025 tinyurl.com/ucfhmfha

Umbrella harms (e.g., Metacognitive) are bolded and represent the total number of instances across all associated sub-harms. Because sub-harms are not mutually exclusive, a single paper may be counted multiple times within an umbrella category if it includes multiple sub-harms with the same type of evidence.

161

Seth Bernstein @sethbern.bsky.social · 19/11/2024

One for HCI

04711