Google combating fake news with AI

Google has developed a project known as the Deepfake Detection Challenge (DFDC) which is a tool that can detect deep fakes which have been generated by artificial intelligence (AI).

Using a dataset of over over 300, 000 videos that have been labelled either real or fake, Google is training a new AI platform that can differentiate the authentic from the counterfeit with far more accuracy than before.

Fake news and misinformation

The dataset itself is a significant step forward in the fight against fake news and false information. Being one of the largest and most diverse datasets related to deep fakes ever devised, it has numerous uses when training models to curb this growing issue.

From defamation to political manipulation, fake news has turned into a veritable whirlwind and has far-ranging and malicious consequences.

As all technology, including deep fakes, becomes more sophisticated and effective, it is now more important than ever to inaugurate tools that can identify such issues with increasing accuracy.

This DFDC dataset is a major step in making deep fakes not only more identifiable but more difficult to create and distribute.

The DFDC dataset

A dataset of such significant importance holds with it several key factors to both be navigated and maximised.

Being so large and diverse, this dataset allows researchers to train far more accurate AI models that display a far more concise method of identification. Additionally, this dataset is open-source, meaning that anyone may use it to develop tools capable of detecting deep fakes.

Along with this, the dataset is also being consistently updated, meaning it will always be adequately equipped to combat all the latest deep fake technology.

Challenges in the dataset’s use

Inevitably with such a new project there are most certainly significant challenges that follow its newfound uses.

Firstly, though its size and diversity are important features, this does make it complex and difficult to use.

Alongside this, the dataset is not perfect by any means and thus the risk of multiple inaccuracies is a high possibility. This may be curbed by having a high-quality labelling process in place to support those who work on the dataset.

New methods to de-fake news

In addition to the innovative creation of this dataset, Google AI is also developing alternative projects to combat deep fake creation.

This includes a separate platform that can detect deep fakes in real-time, bringing numerous capabilities, such as being able to filter out deep fakes from social media platforms before they can gain any traction.

AIBC Asia

The next destination for the industry leading summit is the storied capital of the Philippines, Manila.?AIBC Asia?will be bringing unparallel expertise and a plethora of premium networking opportunities from across the globe to the iGaming and emerging technologies metropolis on the?19th to 22nd of July, 2023.