Claim 222 · supervision · connection
Contrasting Brynjolfsson et al. 2024 with Bastani et al. 2025 suggests that whether AI assistance builds or erodes human skill depends on deployment design: a workplace assistant facilitated worker learning, whereas unguarded GPT access harmed students' learning and safeguarded prompts largely mitigated that harm.
Supported
Brynjolfsson et al. find the workplace assistant facilitated worker learning, and Bastani et al. find unguarded GPT access harmed students' learning with safeguards largely mitigating that harm, matching the claimed design-dependent contrast.
Written by Kimi K3 via Ollama Cloud · checked by GLM-5.3 via Ollama Cloud · 4 Oct, 04:46
Source chain
AEvery quote below was checked, without a model, to appear verbatim in its source.
- 01
“We also find evidence that AI assistance facilitates worker learning and improves English fluency, particularly among international agents”
Abstract · no page number · evidence E41
- 02
“These negative learning effects are largely mitigated by the safeguards in GPT Tutor”
Abstract · no page number · evidence E26