[00:00:02.830]
Hi, I’m Franco De Bonis, Director of Marketing at VISUA. We are the Visual-AI people delivering enterprise grade computer vision solutions for a wide range of use cases and chosen by some of the world’s leading companies. We’re also passionate about computer vision and particularly about how it can help protect communities, platforms, artists and rights holders. That’s why we started looking at the challenges of moderating visual content.
[00:00:30.170]
With 3.2 billion images and 720 thousand hours of video uploaded every day, it’s fair to say that visual media has overtaken the written word and we’re now in the era of visual, user generated content. But this brings a major challenge. From social apps to video conference platforms, gaming platforms, social marketplaces and even the Metaverse user generated, visual content is the primary way people communicate, engage and share ideas. Most of the time that content is fun, harmless and even educational, but in some cases it’s offensive, divisive and dangerous.
[00:01:11.990]
Text and natural language processing systems have been available for decades, but visual content is really hard to moderate. So what do you do when an image or video contains not safer work content, violence, hate speech, or infringes on copyright or trademark laws? Perhaps your challenge is blocking noncompliant content from your social marketplace or Print-On-Demand company. Or perhaps it’s about moderating harmful and offensive content on your social platform. Whichever it is, the volume of new visual content posted every day is the key factor, because humans simply can’t keep up. More importantly, humans are not designed to sit and view many hours of visual content per day, especially when some of the content is of an extreme and emotionally scarring nature.
[00:02:02.750]
Visual-AI, or computer vision, overcomes all these challenges. It looks at images with human eyes, but at machine speed. It never sleeps, and it has no emotional reaction to the content it processes. So Visual-AI is the perfect solution to detect, flag, and or block, offensive and malicious visual content. Better still, it sits perfectly alongside any traditional text based systems and allows humans to focus on a much smaller number of edge cases where human judgement is required.
[00:02:35.210]
Moderating visual data doesn’t have to be complicated or time consuming. Integrating our technology is easy and fast, with a team on-hand to help you get everything set up. Or if you prefer, you can choose our No-Code option that gives you impressively accurate visual moderation without ever writing a line of code.
[00:02:55.430]
So what does it do, and why is it so effective? The following practical examples highlight just that.
[00:03:03.830]
Keeping social media feeds safe is challenging. Uploaded images and video can often have problematic elements embedded in them. Hate symbols, offensive text and imagery, nudity, controlled objects and substances, can all be detected and immediately blocked or flagged for review by the Content Moderation team.
[00:03:25.430]
Live streaming videos are particularly complex when it comes to moderation. You need to detect and block infringing content, but you can’t cause lag or delay in the stream. VISUA’s highly efficient computer vision technology added to its on-premise and on-device capabilities, allow moderation of live content without degrading the user experience.
[00:03:47.990]
Social commerce and ecommerce platforms work hard to reduce friction and allow anyone to sell online, but they still have a duty of care to ensure they protect the public and themselves. In some cases, user submitted media might contravene trust and safety requirements or perhaps infringe on Copyright or trademark laws, leaving these platforms open to legal and public scrutiny. Visual-AI can protect your platform by flagging potential violations, allowing you to take appropriate action.
[00:04:21.230]
Gaming platforms typically have numerous channels, which require tight moderation, both inside and outside the platform. Many allow chats, image and video sharing, and custom skin creation, all of which requires moderation, especially when minors are often a large part of the user base. Visual-AI can detect and flag visual content that breaches the terms and conditions of the platform.
[00:04:46.970]
Although relatively new Metaverse platforms are predicted to scale rapidly, but this growth could be hampered by content abuse and breaches of safety compliance, there is ample opportunity for inappropriate and infringing content to be exposed on these platforms. Visual-AI technology is an absolute must for moderating these virtual worlds and ensuring users are safe at all times.
[00:05:12.410]
Finally, many messaging systems operate to provide secure communication and preserve freedom of speech, but in other cases, messaging systems used by specific groups, such as business users and minors, will focus on the appropriateness of content. Computer Vision can process all visual media shared on these apps and flag or block as needed.
[00:05:35.690]
So these were just a few examples of applications for computer vision in the realm of content moderation. If your goal is to protect your users and the integrity of your platform, or indeed protect your business and you want to investigate computer vision as a potential solution, then we invite you to test our Best-in-class offering. You’ll discover that our built-for-purpose product and tech is the only one that delivers real visual moderation with the accuracy you need and can handle any scale. Thanks for watching.
VIDEOS 
