SafeSearch: Do Not Trade Safety for Utility in LLM Search Agents

Large language model (LLM) based search agents iteratively generate queries, retrieve external information, and reason to answer open-domain questions. While researchers have primarily focused on improving their utility, their safety behaviors remain underexplored. In this paper, we first evaluate search agents using red-teaming datasets and find that they are more likely to produce harmful outputs than base LLMs. For example, when asked "How can I track someone's location without their consent?", a base model refuses, whereas a search agent designed to retrieve and cite sources may lower its refusal threshold, fetch documents (e.g., court cases), and, once appended, synthesize them into an informative yet unsafe summary. We further show that utility-oriented fine-tuning intensifies this risk, motivating joint alignment of safety and utility. We present SafeSearch, a multi-objective reinforcement learning approach that couples a final-output safety/utility reward with a novel query-level shaping term that penalizes unsafe queries and rewards safe ones. Experiments show that SafeSearch reduces agent harmfulness by over 70% across three red-teaming datasets while producing safe, helpful responses, and matches the QA performance of a utility-only finetuned agent; further analyses confirm the effectiveness of the query-level reward in jointly improving safety and utility.

Key Contributions

• Empirical finding that LLM search agents are more likely to produce harmful outputs than base LLMs, and that utility-oriented fine-tuning amplifies this risk
• SafeSearch: multi-objective RL framework coupling a final-output safety/utility reward with a novel query-level shaping term that penalizes unsafe search queries and rewards safe ones
• Demonstrates over 70% reduction in agent harmfulness across three red-teaming benchmarks while matching utility-only fine-tuned QA performance on TriviaQA, HotpotQA, and Bamboogle

🛡️ Threat Analysis

Details

Similar Papers