Facebook has rolled out a new artificial intelligence technology. This is how the company will fight against malicious content.
Facebook announced a new artificial intelligence technology. Its peculiarity is that it can quickly identify malicious content. Thus, the developers hope to make the social network more secure for users. The new AI model uses few-shot-learning. This means that the training model is populated with very little training data. The goal is to reduce the time it takes to detect new types of malicious content from months to weeks.
Malicious content continues to evolve rapidly, and people find new ways to bypass our systems. AI is critically important to evolve with them.
Facebook official statement
It typically takes several months to collect and tag the thousands, if not millions, of examples required to train each individual AI system to discover a new type of content.
The new technology recognizes over a hundred languages and works with both images and text. Fast Learning AI is designed to complement existing methods for assessing and removing malicious content. The system has already been tested to detect harmful misinformation about COVID-19 vaccination.