Researchers have developed a new approach to AI security that employs text prompts to better protect AI systems from cyber threats.
This method focuses on the creation of adversarial examples to protect AI security and prevent it from being misled by inputs that are typically undetectable to humans.
The prompt-based technique streamlines the generation of these adversarial inputs, allowing for quicker response to potential threats without extensive computations.
Preliminary testing has shown that this method can effectively safeguard AI responses with minimal direct interaction with the AI systems.
The research, ‘A prompt-based approach to adversarial example generation and robustness enhancement,’ is published in Frontiers of Computer Science.
How can prompts prevent cyber attacks?
Dr Feifei Ma, the lead researcher, outlined the process: “Our approach involved initially crafting malicious prompts to identify vulnerabilities in AI models.
“Following this identification, these prompts were utilised as training data, enhancing AI security by resisting similar cyber attacks in the future.”
Malicious prompt texts were first constructed for inputs, and a pre-trained language model can generate adversarial examples for victim models via mask filling.
Models trained with adversarial prompts were less likely to succumb to similar attacks, demonstrating an enhancement in their defensive capabilities.
Enhancing AI security across key sectors
Subsequent experiments indicated that this training approach improved the robustness of AI systems.
“This method allows us to expose and then mitigate vulnerabilities in AI models, which is especially critical in sectors like finance and healthcare,” said Dr Ma.
The research indicates that AI systems trained with these adversarial prompts are more capable of resisting similar manipulation tactics in the future.
This could potentially improve AI security against cyber threats in several key industries.
