VISUA Answers 7 Common Questions About Online Content Safety

There is an enormous amount of content in the online space and some of it is created in an intentionally malicious manner. This is why academics have noted that, “Malicious web content is a serious problem on the Internet today”. Avoiding this sort of content can be a nuisance while navigating the internet. That is why many online platforms and online businesses use various tools and techniques to mitigate the amount of unwanted content on their platforms. The effectiveness of these approaches vary which is why some sites have a greater degree of malicious content than others. VISUA has a Computer Vision API (also known as Visual-AI) which has been trained to detect and process a variety of visual media to solve these problems in an automated and efficient manner. Newstrail journalist Ethan Frost recently interviewed Franco De Bonis, the marketing director of VISUA. He provided answers to seven common questions about online content and how to manage it.

Newstrail: How can a Company Prevent Phishing? 

De Bonis: The more technology companies try to prevent phishing with new technologies, the more phishers will innovate their own techniques. It has always been important for us at VISUA to use our technology to solve the most pressing issues facing our day-to-day lives, so we wanted to break this cycle by using a Visual-AI to develop an effective system that anti-phishing platforms could deploy to help improve visual threat detection. The AI looks at emails and web pages like a human does rather than looking at the code behind it and so provides an early warning when it detects high-risk brands and other visual factors in emails and websites. We have developed a proprietary technology stack which provides real-time analysis and detection. 

Newstrail: What is the Best Way to Monitor a Brand?

De Bonis: Most of the content produced today is visual in nature, whether images or video. But most brand monitoring platforms still employ systems that track brand mentions through text alone. This produces a far less comprehensive picture of how consumers are engaging with brands. To solve this issue, VISUA has developed a complete technological solution which is focused on identifying visual brand mentions at a scale and level of precision which the market demands. We can also detect key objects and also text that is burnt into the visual media, so we don’t just deliver impressions, we also deliver the context and other key signals that go with that impression.

Newstrail: How can IP Owners Monitor Copyright and Trademark Compliance?

De Bonis: People and companies who own Intellectual Property like copyrights and trademarks naturally want to protect them from fraud and misuse and want to ensure that their property is adequately protected. Additionally, online platforms want to ensure that the products which they sell are compliant with IP law. But the scale of content which needs to be evaluated is far too great for humans to be able to efficiently process. This is why VISUA developed a dedicated technology stack that is focused on delivering visual copyright and trademark protection. We can not only detect exact matches of logos and marks but also close matches. This is especially useful where brand assets are modified to pass-off as the genuine brand or creatively adapted to fit a different theme.

Newstrail: How can an Ecommerce Platform Detect Counterfeits?

De Bonis: Counterfeiters know that ecommerce platforms and marketplaces have text detection capabilities which would flag their fake products. So they use visuals to disguise their counterfeited products.. A good example is a listing that might say ‘Women’s Handbag’ but when you look at the image it is clearly a Louis Vuitton bag and might have more descriptive text embedded in the image. Ecommerce platforms need a tool that is capable of analyzing large amounts of visual data to stop counterfeit goods from reaching their storefronts. VISUA has built a unique technology that is focused on solving the problem of counterfeit goods. 

Newstrail: Similarly, How can a Business Authenticate Products? 

De Bonis: Yes, Product authentication is to the supply chain what counterfeit detection is to online. The rise in ecommerce has led to a rise in the production and sale of counterfeit goods. Brands want to ensure that their goods are genuine and protected at all levels of the supply chain, from warehouse staff, to customs officers, all the way through to the consumer themselves. It is for this reason we have built technology which is specifically architected to deliver fast, accurate and simple hologram authentication by anyone. Our Visual-AI system uses consumer devices and requires no training to implement.  

Newstrail: How can a Brand Monitor the Success of Sponsorships? 

De Bonis: I think It is important for companies to monitor their sponsorship value and be provided with ROI data. To present this information effectively requires going beyond the traditional channels and delivering important insights in real-time. We know that partial data sampling from traditional broadcast media is no longer viable. In order to be able to accurately calculate sponsorship ROI a platform needs all the data from all the channels and all media types. But at its core, and this is true of any campaign measurement tool, brands must have precise impression data that they can trust. The Visual-AI which we have developed can measure broadcast, online, social and even print impressions from visual media using the same methodology. This allows companies to process any stream in real-time which delivers instant insights to a platform. As a result, companies can compare the impact of different channels without the need to question the reliability or accuracy of their data. 

Newstrail: How can a Platform Moderate Visual Content? 

De Bonis: The vast majority of content which is generated by users contains visual media. This includes platforms from social media to messaging apps and even the metaverse. Because of the massive volume of content which users generate on a daily basis, human moderation alone cannot cope, and this means that platforms wanting to moderate their content effectively will need VIsual-AI Solutions. VISUA has developed a full service technology that uses several computer vision technologies to accurately moderate visual media at the scale   

VISUA Media Interview

Franco De Bonis spoke with journalist Ethan Frost about the European Union’s Digital Services Act and how Visual-AI can help in meeting and maintaining compliance:

Final Thoughts 

There are many issues facing online platforms which are exacerbated by the substantial amount of content which they process on a daily basis. The volume of content which they are required to handle is far too large to be monitored by humans. Many AI solutions to these problems are not able to accurately process visual information which comprises a great deal of the content platforms are working with. As a result VISUA has developed a Visual-AI which can process content at a speed and accuracy which makes the process incredibly efficient. This technology has a number of use cases for businesses operating in the online space.