GBTB

Scam Alec

Comment

Inspiration

The inspiration behind Scam Alec stems from the alarming rise in financial fraud. The public often thinks that they are not that "stupid" to be scammed however scams are often more real than they think. Hence we thought of using humour to gain attention.

What it does

Our hack consists of 2 portions, a telegram bot and a telegram channel. These tools are designed to tackle the escalating threat of financial fraud by enhancing public awareness, providing real-time information, and educating users on how to recognise and respond to scam attempts.

We have chosen to use telegram as the main platform as more than 50% of scam victims are young adults aged 20-39. We believe that telegram is a common platform that young adults use. In fact, a survey conducted revealed that the largest share of telegram users as of 2023 was between 25 and 34 years of age. The prevalence of telegram among young adults means that these resources will be accessible for them with a low barrier to entry, via both smartphones and laptops.

  1. Telegram Bot @scamalec_bot Check Scam Feature: Users can send suspicious text messages to the bot to check how likely they are to be scams. This empowers users with immediate feedback and helps them identify potential scams before falling victim. Real Stories: The bot provides access to real stories of individuals who have been scammed, raising awareness through personal experiences and educating users on common scam tactics. Latest News: Users can stay updated with the latest anti-scam news, helping them stay informed about new scam techniques and prevention strategies.

  2. Telegram channel @scamalec This channel aims to engage the public using engaging and light-hearted headlines as a primary tool. Each message will feature an attention-grabbing headline designed to encourage clicks on the story link. Daily updates will be provided to serve as regular reminders of the prevalence of scams, emphasizing the importance of staying vigilant.This channel will use humour as a primary tool to entice the public. Funny/humourous phrases will be the header of each message which aims to draw their attention to click onto the link of the story. There will be daily updates, which can serve as a daily reminder that getting scammed is more common than they think.

Overall, the use of stories and news aims to bridge the gap in public awareness against the ever evolving scam tactics. By having light-hearted and attention-grabbing news and headlines to grab the public’s awareness, we hope that they will be looking forward to the daily updates which can offer a variety of scam tactics. Furthermore, the use of machine learning in the check scam function will mean that there is opportunity to continuously update the dataset/model to recognise new and evolving scam patterns which ensures that the bot stays effective even as scammers change their tactics.

How we built it

Our hack is a Telegram channel and bot. The primary programming language used to implement our Telegram bot is Python. The Telegram channel was created using the existing Telegram GUI. Implementation of the telegram bot functions: The checkscam function uses a pre-trained model. A text classification pipeline (ScamLLM) using the ‘transformers’ library was initialised.

The news and real stories function was done using web scraping from the website https://www.scamalert.sg, which is a website by The National Crime Prevention Council (NCPC). NCPC is a non-profit organisation committed to promoting public awareness and concern about crime and to propagate the concept of self-help in crime prevention in Singapore, and hence the information on the website is credible and relevant.

  • The retrieved HTML content is parsed using BeautifulSoup, a Python library, which makes it easy to navigate through the HTML tree structure.
  • To extract relevant data, specifically latest scam-related news and real stories, the respective functions search for specific HTML elements (such as divs with particular classes) that contain the news articles. For each article and real story, it extracts key information including the date, title, and link.
  • To display the news or stories, when a user issues the /news or /realstories command, the bot displays the latest news and real stories respectively by cycling through the stored list, providing navigation buttons for "Next" and "Previous" to view different articles or stories.
  • The integration of these web scraping functions enables the bot to dynamically fetch and display up-to-date content from external sources, ensuring users have access to the latest information on scams and real-life experiences.

Challenges we ran into

One of the main challenges we encountered was ensuring the accuracy and relevance of the scraped data. Additionally, integrating the machine learning model for text classification with the Telegram bot posed challenges related to model size and deployment constraints. Ensuring the bot remains responsive and efficient while performing these operations was another significant challenge.

Accomplishments that we're proud of

We are proud of successfully integrating a machine learning model into our Telegram bot to provide real-time scam detection and classification. Additionally, our web scraping implementation dynamically fetches and displays the latest scam-related news and stories, offering valuable information to users. We are proud of creating an engaging and interactive platform that raises public awareness and helps users stay informed about scams.

What we learned

  • Build a telegram bot
  • Web scraping
  • ML model integration into telegram bot

What's next for GBTB

As we have created a Telegram channel with our bot to implement our solution, we have thought of ways to enhance our solution to make it even more comprehensive. For instance, currently our channel sends daily article reads to our channel subscribers. Apart from article reads, we were considering to also include short quizzes or interesting games in our channel. This could further increase our subscribers' engagement, enhancing their critical thinking in scam detection and alertness. Moving forward, GBTB intends to enhance our solution's interactivity and effectiveness to make it an adequate solution in helping users to easily yet accurately detect financial frauds

Built With

Share this project:

Updates