The Center for Data Innovation spoke with Nageela Yusuf, CEO and founder of TikBox, an online platform that simplifies the legal workflow and metadata tagging for copyright and provenance assertions on various digital files to better support creatives in the new era of large language models. Yusuf discussed how the changing technological landscape is affecting the creative industry, the need for searchable copyrighted metadata on the Internet, and how we can ensure the voices of the creative industry are kept relevant in AI development.
This interview has been edited.
Ayesha Bhatti: What is your background, and why did you decide to found TikBox?
Nageela Yusuf: I’m a trained photographer and videographer turned technologist and I have been working in media tech for over 15 years, asserting copyrights on commercial projects. My team and I previously built the Cerebriam Studio online editor to create impactful video compositions in multiple aspect ratios. Our work sits at the crux of creative and automated, and we do a lot of work with machine learning, specifically with computer vision and generative AI.
Given our technical experience and creative background, we were amongst the first to start asking questions about the commercial use of assets for machine learning purposes that did not engage copyright holders meaningfully, as well as identifying a solution to this very modern problem of copyright assertion and provenance labelling. For us, copyright assertion and provenance labelling are different sides of the same coin. This is where TikBox comes in, dealing with both to offer a streamlined process for creatives online.
TikBox offers the opportunity to find the balance between protecting creative works and supporting the development of generative AI. It’s incredibly important that we get things right, and with the slew of court cases at the moment investigating this new relationship between developers and copyright holders, as well as the increased cyber security threat of disinformation from generative AI, we want to protect people from malicious content, and ensure that the creators behind the vast amounts of content used for training data have their voices heard.
Bhatti: How exactly does TikBox support creatives?
Yusuf: TikBox puts content creators, whether they are an enterprise organisation or an independent artist, at the front and centre of how their assets are used. Copyright and provenance assertion, as well as licensing for creative content needs to be modernised for the age of generative AI. Given the widespread scraping of the web, and ease with which assets can be manipulated, it’s important that creators have tools to distinguish their original content, as well as the agency to decide how this content can be used for commercial purposes, and finally the mechanism to create agreements and automatically collect licensing fees. We’ve simplified this entire workflow to the point of ticking a few boxes—it’s this simplicity that has inspired the product name.
There is immense value in the work of creators, not only for traditional areas of licensing such as advertising, merchandising, entertainment, and media, but now also for training AI models. TikBox is evolving to meet both the current and future needs of creators and licensors in a changing technological landscape.
Bhatti: How important are meta tags in the world of Internet data?
Yusuf: Meta tags have always been central to how content is displayed and handled by search engines and browsers. They are particularly valuable in today’s age, when searching for and scraping data from the internet is commonplace. Metadata has also been with us for a long time and essentially provides information about the file. Traditionally, it has carried technical information necessary for the playback or reading of digital files. With the advent of commercial digital music and photography, copyright information was added. Copyright related metadata should be handled in a similar way to meta tags, in that it should be searchable.
Bhatti: What is the difference between TikBox and the C2PA initiative?
Yusuf: The C2PA standard is an interesting initiative founded by Adobe, the BBC, and Microsoft to standardise the labelling of digital content. This allows users to clearly understand the origin of the content and any attempts at manipulation. It provides a generic framework available as a software development kit for use in various products. The BBC and Adobe have already implemented it in their respective products. TikBox has also implemented C2PA. For us, this is a small part of a larger product workflow designed to streamline the end-to-end process of asserting rights and provenance, as well as simplifying commercialisation. TikBox’s workflow has been developed to create custom licences, add copyright and provenance meta data to the assets in question, and conclude with payment upon completion of a licensing arrangement. Our integration of C2PA introduces an additional layer of security and enhances the interoperability of copyright and provenance data.
Bhatti: How do you envision TikBox evolving in the next few years, and what are your long-term goals?
Yusuf: It’s an exciting time to be innovating at TikBox. For many of us with exposure to the creative sector, generative AI, and the regulatory environment, it’s obvious that changes across intellectual property, privacy, and security are underway. At TikBox, we are committed to supporting creators, licensors, and those enjoying trustworthy content, enabling them to do so on the web, on their mobile devices, and in every place where they create and consume content. As the months go by, we expect to see refinement in how our product is used across various industries. One thing we can be certain of is that issues surrounding intellectual property, privacy, and security will impact every industry.
