WorldPolitics

When chatbots become ‘answer engines’: AI, geopolitics and truth

Read full article at source ↗

Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 04/11/2026, 14:30:25 UTC.

Briefings

Research from 2025 at Hibrid.info found that AI chatbots like ChatGPT and DeepSeek Chat often produce factually accurate but misleading answers on geopolitical topics, while Alice shows higher refusal rates and ideological deviations. The study tested 100 prompts across Kosovo and six Western Balkan nations, revealing that location significantly alters how chatbots respond, sometimes switching to Russian or refusing to answer entirely.

Key points

• Three AI chatbots (ChatGPT, DeepSeek Chat, Alice) were tested on 100 standardized prompts regarding Kosovo and regional history.
• ChatGPT and DeepSeek Chat showed high accuracy but produced similar, often misleading narratives on sensitive geopolitical topics.

Context

The study examines how AI chatbots function as reference tools in the Western Balkans, where political narratives are contested. Findings from the 2025 research report 'The Chatbot Version of Truth' suggest that plausibility does not guarantee truth, especially for sensitive topics.

Limitations

Data covers only 100 prompts across six Western Balkan nations; Alice showed higher refusal rates and ideological deviations; some responses were mixed languages or refused entirely.

Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 08/11/2026, 01:15:03 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.

View archived briefing on IPFS
bafybeighwa2m2puxvplpkgiod3rzynutxpvbc353r5y555maspkxi6ipv4