ScamSense - Daily Financial Fraud and Scam Alerts

Scam Sense is an Alexa skill that empowers users to stay informed about financial fraud and scams. Every morning, it fetches the latest articles and uses AI to generate concise daily summaries.

Comment

A blue banner that reads "Enable on your smart speaker or mobile app, click here!"

Inspiration

The idea for Scam Sense emerged from the increasing prevalence of financial fraud and scams that impact individuals and communities worldwide. Phishing and scam activity has increased 94% since 2020. In an age where information is key, staying updated about the latest threats can be a daunting task, especially for those without financial expertise. Scam Sense aims to bridge this gap by delivering concise, daily news summaries and scam avoidance tips though an Amazon Alexa Skill about scams and fraud, empowering users with the knowledge they need to protect themselves and others. Alexa was chosen as a target platform due to its ease of use, particularly for senior citizens since 2 out of 5 seniors have been victims of a scam. Between 2019 and 2020, the percentage of seniors owning voice-activated home assistant devices (like Amazon's Echo) rose from 17% to 29%. By being able to request reports via simple voice and text commands, ScamSense can empower users to protect themselves.

What it does

Sample response from ScamSense

Scam Sense is an Alexa skill that provides daily updates on financial fraud and scam-related news. By integrating with The News API, the skill fetches relevant top stories, summarizes them using advanced AI (using Llama hosted on SambaNova Cloud), and stores the results in a DynamoDB database. Users can then access these summaries through their Alexa-enabled devices, receiving up-to-date information in a clear and concise manner.

How we built it

Scam Sense was built using a combination of Amazon Web Services (AWS) services and third-party APIs:

  • News Aggregation: Top stories for the day are fetched using The News API.
  • AI Summarization: SambaNova Cloud's hosted Llama API programatically generates concise, user-friendly summaries and scam avoidance tips using the fetched articles.
  • AWS Infrastructure: The architecture includes Lambda functions for serverless compute, DynamoDB for data storage, and EventBridge to automate daily updates.
  • Alexa Skill Integration: The Alexa skill is linked to the backend, enabling seamless user interaction. The project was developed using Python for the backend logic, and AWS CloudFormation was used to simplify deployment.

Challenges we ran into

  1. Data Filtering: Ensuring the relevance of fetched articles required fine-tuning API queries and filtering mechanisms. It was important to ensure all information included in the ScamSense reports was related to financial fraud and crime.
  2. Summarization Accuracy: Generating concise and accurate summaries using AI posed challenges, which was resolved via prompt engineering. Additional prompt engineering was required to add scam avoidance tips and to produce output that is similar to a radio broadcast for consumption when Alexa reads the daily scam and fraud report aloud.
  3. Alexa Skill Integration: Configuring the skill to interact smoothly with the backend and ensuring real-time responsiveness required extensive testing.

Accomplishments that we're proud of

  • Successfully integrating advanced AI for summarization into an Alexa skill.
  • Creating a fully functional application that delivers value by spreading awareness and avoidance tips about financial fraud.
  • Developing a robust AWS infrastructure to ensure scalability and reliability. This architecture is economical (leverages pre-computed responses and maximizes use of Amazon's free tier) and can be reused for similar educational Alexa skills by simply modifying the AI prompt or Lambda API request.

What we learned

  • The importance of clear and concise communication when dealing with complex topics like financial fraud.
  • How to leverage AWS services for rapid development and deployment.
  • Effective integration of third-party APIs with custom backend solutions.
  • Insights into designing user-centric Alexa skills for practical applications.

What's next for ScamSense - Daily Financial Fraud and Scam Alerts

  • Localization: Expanding support for multiple languages to reach a broader audience.
  • Advanced Insights: Adding features like trends and statistics on financial fraud cases.
  • Community Engagement: Partnering with financial institutions and advocacy groups to spread awareness further by including their information in ScamSense reports.

Built With

Share this project:

Updates