AI Chatbot Security Assessment- CYbersecurity

₹600-1500 INR

Completed

Posted 29 days ago

₹600-1500 INR

Paid on delivery

You have been provided with the following information about a project that a team within your organisation wants to progress. Please prepare a presentation (20- 25 mins) outlining of the process you would follow in order to assess and improve the security of this project as a whole. Please focus your attention on the following: What sort of information will you need to progress your assessment and how would you obtain it? What are the key aspects of this project that need attention and how can they be addressed? What methodologies would you apply in assessing the security risks, governance, compliance? How would you recommend risks and vulnerabilities are addressed by the project? Scenario The organisation has recently proposed a new project – a generative AI driven chat bot. The purpose of this project is to enable customers to change the address related to their policies, through an intelligent AI driven support agent. The following is a basic diagram showing how the proposed multi agent system is to be deployed. The system initiates the orchestrator agent on signal that the user wants to change their address. The orchestrator manages the flow of the interaction, and decides which sub-agents to invoke in order to answer the users query. The database agent generates SQL queries against the customer database, updates addresses and returns the current user address. The Report and Formatter agent summarises the information given to it, formats it into an acceptable output and returns that to the users chat session. The underlying host infrastructure is an enterprise level setup and you can consider that general security best practice is surrounding this system. The back-end LLM is a frontier model utilized through a third party API call to the provider. NOTE - i want to create a PPT for this to describe in highly technical use case with showing the best security solution using aws cloud for example..like how you will secure each point .pls see the attached screenshot

Project ID: 40383084

About the project

5 proposals

Remote project

Active 29 days ago

Looking to make some money?

Email address

Benefits of bidding on Freelancer

Set your budget and timeframe

Get paid for your work

Outline your proposal

It's free to sign up and bid on jobs

Awarded to:

@manoval09

Hi, Your project description cuts off mid-sentence, but the core need is clear — you want a security assessment of an AI chatbot, and the real attack surface for LLM-based systems goes well beyond standard web vulnerabilities into prompt injection, insecure output handling, and context poisoning. My approach follows the OWASP LLM Top 10 as the baseline, using `garak` (an open-source LLM vulnerability scanner) and `PyRIT` (Microsoft's AI red-teaming toolkit) to probe the chatbot systematically. I'd prioritize the highest-impact vectors first: direct and indirect prompt injection, jailbreaking via role-play framing, PII leakage through model completions, and over-reliance on unvalidated retrieved context if RAG is involved. Within the $600 scope, I'd deliver a tiered report — critical and high findings with proof-of-concept examples and concrete remediation steps, not just a checkbox list. Before I lock in the scope: is this a cloud-hosted chatbot built on an API wrapper (OpenAI, Claude, etc.) or a self-hosted fine-tuned model? That single answer determines which attack surfaces dominate and how I'd allocate testing hours. Best regards, Val

₹1,500 INR in 7 days

0.0

(0 reviews)

0.0

5 freelancers are bidding on average ₹5,210 INR for this job

@SmartCodeDev

Hello I can create a 15–20 min technical PPT explaining how to assess and secure your GenAI chatbot using AWS architecture. It will cover threat modeling (STRIDE, OWASP LLM risks), required system information, and key security risks in each agent. I will also show AWS-based protections like IAM least privilege, VPC isolation, WAF, Secrets Manager, CloudTrail, and GuardDuty. Special focus will be on securing the orchestrator, database agent, and preventing prompt injection or unsafe SQL execution. The final PPT will be clean, structured, and ready for technical presentation. Ready to start immediately.

₹12,000 INR in 7 days

5.0

(2 reviews)

1.8

@PetrusLubbe

‼️ONLY PAY WHEN YOU'RE 100% HAPPY‼️ I see your project focuses on a generative AI chatbot that updates customer policy addresses—a system with complex multi-agent orchestration and sensitive data handling. Security here hinges on safeguarding data flow across agents and securing third-party API calls to the LLM provider. My approach would involve mapping data access points, applying threat modeling, and leveraging AWS security tools such as IAM for strict access controls, VPC for network isolation, and AWS Shield for DDoS protection. I'll design practical controls to secure database queries and API interactions without sacrificing performance. While I’m new to Freelancer, I’ve done similar API security and AI system hardening off-platform with measurable risk reduction. Let’s chat! Worst case, you get a free consultation and real insight. Regards Pietie Lubbe

₹1,050 INR in 14 days

0.0

(0 reviews)

0.0

@flapjakc

Hello, I’ve built and delivered similar systems, so I can implement your requirements quickly and cleanly. I’ve assessed AI chatbot architectures with focus on LLM risks, data security, and AWS-based protection layers. My approach: map data flows, identify attack surfaces (prompt injection, API abuse, data leakage), then apply STRIDE/NIST frameworks with governance and compliance checks. I’ll design a secure AWS architecture using IAM, VPC isolation, API Gateway, WAF, encryption, and monitoring, and present it in a clear 20–25 min PPT. Timeline: 2–3 days. Let’s start.

₹10,000 INR in 7 days