EcoSort Security Solutions

• 

  Sorting with AI

• 

  Step 1 - Item image or file

• 

  Step 2 - Examples of Prompt Injection (Outcome)

• 

  Step 3 - Examples of Code Injection

Inspiration

In today's digital age, safeguarding user data and ensuring mobile game app security is more critical than ever. Most people don't think about sustainability in their day to day lives, most people just throw away waste in whatever is convenient for them. If we gamify recycling and incentivize it by partnering up with local businesses to provide prizes, we promote sustainability. Yet, challenges in data protection, integrity, and user safety remain, requiring robust solutions.

What it does

EcoSort Security Solutions is at the forefront of protecting user information and preserving game integrity from cyber threats. Hackers exploit vulnerabilities like image manipulation to misclassify recycling items, which may seem trivial but holds global implications. Such manipulations could devastate other business industries.

How we built it

We will use Python to develop a game that utilizes your mobile camera for classifying recycling items using AI. This innovative approach ensures accurate categorization. However, our focus extends beyond functionality to safeguarding against potential exploits, such as prompt injection engineering. By understanding these vulnerabilities, we're proactively defending game apps and broader systems against cyber threats. We can show how prompt injection engineering can manipulate the outcome of the prompt which can be a file or image. From this standpoint, we can see how hackers can easily exploit game apps and other systems.

Challenges we ran into

Time- Trying to build a simple and functional app is really difficult with the amount of time that we have Concept development Learning how to use the API The code generated by ChatGPT doesn't work with the current version of the API Some APIs require permissions or API keys Have to purchase a subscription in order to use more features

Accomplishments that we're proud of

Though our game app prototype isn't yet functional, we've demonstrated our concept through mock examples in the LLM. By showcasing prompt injection engineering, we've highlighted one potential manipulation method. Given more time, we could delve deeper into various cybersecurity threat strategies, showcasing our commitment to comprehensive protection in game app development.

What we learned

Some of the team members didn't know that LLM such as ChatGPT and Claude AI can easily classify recycle items by simply uploading an image of the item and using a simple prompt. We learned that AI has proven to be a powerful tool for creating mock examples of vulnerabilities, which can be invaluable for cybersecurity firms. We learned about prompt injection engineering.

What's next for EcoSort Security Solutions

We will complete the prototype for our game app and fortify it against potential vulnerabilities. Utilizing advanced AI and Python, users can accurately classify recycling items via their mobile cameras, ensuring data integrity. Yet, our vigilance extends beyond image manipulation to encompass threats like patching vulnerabilities and guarding against scams and phishing attempts impersonating the game. By addressing these risks, we're dedicated to delivering a secure and enjoyable gaming experience with robust safeguards in place.

Built With

• chatgpt
• claudeai
• openaiapi
• python