Address chatbot-induced challenges computationally and pedagogically, including tone safeguards, bias reduction, and user-side prompt engineering¶

Strategy · All strategies
Evidence · 3 claims (3 for) · 2 studies (2 design), q2 · 0 of 2 report an effect size · 3 claims rest on one study

Description¶

The article recommends addressing the chatbot's challenges both computationallally and pedagogically: safeguards and careful fine-tuning can reduce bias and tone problems, and users can be guided to personalize interaction via prompt engineering. The authors note lengthy posts (averaging 298 words versus users' 79) may overwhelm some users and suggest prompts such as "limit your answer to 100 words" or hyperparameter configuration.

Design Implications¶

Context¶

Requirements¶

• Further computational work (e.g., safeguards) and educational work, including reflection on the interaction itself.

Constraints¶

• The authors hypothesize the negative tone stems from the fine-tuning data, so tone problems can recur without intervention; alignment attempts did not always succeed.

Target Learners¶

• Users of LLM-based dialogic chatbots on contentious topics

Target Learning Goals¶

• Sustaining a safe dialogic space while limiting bias, negative tone, and post length

Claims¶

• LlamaLo exhibited bias and presented contested assertions as facts, notably in the Israeli-Palestinian conflict discussion [+W]
• LlamaLo occasionally employed an aggressive or sarcastic tone, sometimes without user provocation, though it changed tone after being commented on [+W]
• Fine-tuning frontier LLMs on authentic tutoring data improves performance on all six conversational and pedagogical benchmarks, moving values toward human-tutor levels [+W]

Related Strategies¶

Examples¶

-

Key Sources¶

• Yifat Ben-David Kolikant, Omri Hadar, Asaf Salman. (2025). Talk to the machine: Unleashing the potential of AI to scale dialogic education and reduce polarization. Intern. J. Comput.-Support. Collab. Learn. https://doi.org/10.1007/s11412-025-09461-8