Dive deeper

Problem statement

The technology that is used to create Deep Fakes and Synthetic Media is the same. However, the problem is that DeepFakes are deployed to cause harm.

Although sophisticated DeepFake detection tools exist , there is no free, easy-to-use tool that is widely used and looks at video, audio and text simultaneously.

If we share, comment or engage with DeepFake content unknowingly across different social media platforms, we feed the algorithm and contribute to spreading harm.

Therefore , we believe it should be easy and effortless for everyone to verify authentic digital content in our daily lives.

If we succeed in developing an innovative way to confront DeepFakes online we could help everyone navigate digital content with more competence.

Challenge

How can we help tech-aware people to distinguish real from fake online content by empowering them with effortless independent tools to spot deepfakes (image, video, audio, text) which pollute or cause harm in their information consumption and especially when there is uncertainty such as Covid-19, political elections, war, economic recession, etc. instead of being manipulated by or unintentionally perpetuating misinformation?

Sub-questions

• How can we design a system of rewards and incentives for sharing accurate information online?
• How can we inspire tech-savvy people to use the deep fake alert tool frequently?
• How can we create a data visualization function for user feedback and reflection?

Read the article here

Criteria:

• The solution should be easy to employ, and not include unnecessary extra steps in the user experience (e.g. integrated)
• The solution should be focused on audio and/or text analysis and if it includes video/photo -which is already being widely addressed- it should bring novelty to the design
• The solution should be open-source and GDPR compliant
• The solution should be mobile device compatible

Summery

The goal of the DeepFake Alert solution is to help people to discern what is fake or real and what is original or manipulated on the Internet. The tool should be open source and easy to use. It will empower individuals who seek to be well informed to navigate with greater digital fluency.

Additionally, it should have a function that allows users to design personalized tracking and monitoring analytics using data visualizations that are easy to share in a community of well informed fellow citizens.

For this challenge in particular, Find Out Why partnered with Microsoft and its director of Technology Operations Ashish Jaiman, who is a thought leader in the role of DeepFakes in disinformation and how AI is the solution to the right direction.

Sources for open data

• Deepfake Detection Challenge Dataset
• Deep Fake Image detection using Anchored Pairwise Learning Approach
• DeepFake-o-meter: An Open Platform for DeepFake Detection
• The world's first detection platform for deep fakes