Overview
This service focuses on evaluating the security of AI systems deployed within agrifood digital platforms. The objective is to detect risks such as insecure data flows, improper access control, and unintended exposure of internal information. The assessment involves a structured review of how the AI system interacts with external components, users, and data sources. A combination of automated tools and guided analysis is used to identify potential vulnerabilities or misconfigurations. This helps ensure that AI-based systems are deployed securely and responsibly, in line with the expected reliability and trustworthiness standards of the agrifood sector.
More about the service
How can the service help you?
AI systems can process sensitive data, interact with users, and integrate with third-party services, which may introduce security risks or unintended data exposure. Prior to the assessment, these issues might remain undetected. After the service, the customer receives a detailed report identifying security weaknesses, their severity, and actionable recommendations for mitigation. This contributes to improving the robustness and resilience, transparency, and market readiness of the solution.
How the service will be delivered
The service is delivered remotely. The customer must provide access to a working version or testing environment of the system (URL, test credentials, API endpoints), along with a brief technical description of the AI system to be evaluated.
The deliverable is a structured report including identified risks and vulnerabilities, severity classification, and recommended mitigation strategies.
Meetings may be scheduled during the evaluation process to clarify functional aspects or address specific questions. A follow-up session can also be arranged upon request to discuss the results and support decision-making.
There are no restrictions related to time of year; the service can be provided at any moment.
Service customisation
The service can be tailored to different types of AI systems, focusing on their interaction with users, external services, and data sources. The analysis combines the use of automated tools and manual inspection to identify weaknesses related to access control, input handling, and information disclosure. It is applicable to systems with a functional version or test environment, and it follows a structured, non-intrusive methodology aligned with secure-by-design principles.
Sector
Aquaculture
Arable farming
Forestry
Food processing
Greenhouse
Horticulture
Livestock farming
Tree Crops
Viticulture
Test type
Data
Design / Documentation
Software or AI model
Type of service
Cybersecurity
Data analysis
Desk assessment
ELSA assessment
Performance evaluation
Test design
Country of delivery
Country of delivery:
Remote
Service provider(s):
-
University of Cordoba (UCO)
| Website