Written evidence submitted by Yoti (SMH0039)
About Yoti
Yoti is a digital identity company that makes it safer for people to prove who they are. Founded in April 2014, we started by creating a secure Digital ID app which gives people a safer and instant way to prove their identity, with no need to show identity documents or share an excessive amount of personal data. Yoti now provides verification solutions across the globe, spanning identity verification, age verification, age estimation, eSigning and authentication. We’re a team of over 400 people, working together to shape the future of digital identity.
We’re committed to making the digital world safer for everyone. Our seven ethical principles guide us in everything we do and we’re held accountable by our independent Guardian Council, whose minutes we publish. With an award-winning social purpose strategy, we’re always looking for new ways to explore what (digital) identity means globally. The journey isn’t one we’re making alone, but with the help of policy advisers, think tanks, researchers, humanitarian bodies and everyday people.
● Transforming the way individuals can prove their age and identity
● Increasing security and privacy of personal data
● Helping to create age-appropriate experiences and safer communities online
● Creating the most reliable and comprehensive identity verification solutions
● Shaking up the way we sign documents
Technology as a force for good - Yoti was founded on seven business principles which guide our actions. Yoti is also a founding UK B Corp meaning we aim to balance profit with purpose.
Security credentials - We commission regular external audits of our business and have been certified to meet some of the world’s most stringent security standards, such as ISO 27001 and SOC2 Type II. We are also certified by the UK Government under the UKDIATF.
A transparent, open and honest approach - Yoti publishes regular white papers to build trust and understanding of our technology.
Response
Social media and search engine business models often seek to stimulate user engagement to maximise ad revenue. Therefore, algorithms are often designed to optimise user engagement. This means they will tend to amplify emotional, controversial or polarising content, which can help propagate harmful or false information.
A solution of addressing these issues could be to find the right balance between user retention and engagement on a given platform, and the way that content is presented to users. This could include introducing requirements on algorithm transparency, and accountability frameworks.
Ranking algorithms rely on users’ data to predict their preference, and offer content that will enhance their engagement. As we have said above, they automatically will present content that can be considered sensationalist, and therefore can lead to the proliferation of misinformation.
Alghoritimic biases can also create and exacerbate echo chambers, in which users find themselves locked. To mitigate this, platforms could give users a greater degree of choice over algorithmic recommendations. Algorithms themselves could also be subject to regulatory oversight, and be assessed by regulator staff and research.
Generative AI and large language models enable the rapid production and sharing of seemingly realistic but harmful or false information. This includes deepfake or false news articles.
Regulation should aim to encourage the production and adoption of safeguards such as digital watermarks, verified content uploads using second e-signatures. Solutions already exist on the market today, and Yoti offers a number of robust detection technologies like MyFace Live and Secure Image Capture (SICAP) solutions. More information can be found here: https://developers.yoti.com/age-estimation/secure-image-capture.
How effective is the UK's regulatory and legislative framework on tackling these issues?
A key part of the UK’s regulatory and legislative framework is the Online Safety Act. The Act can be seen as a strong starting point, but it has become obvious in recent months that it must evolve quickly to address the risks posed by generative AI and algorithms.
We would reiterate our call for greater collaboration between digital regulators (especially the ICO and Ofcom), to ensure that there is holistic oversight of content and its spread.
How effective will the Online Safety Act be in combatting harmful social media content?
The Online Safety Act should be seen as a positive first step, setting the foundation for addressing harmful content. However, it could be more specific regarding generative AI and algorithmic oversight.
We have repeatedly highlighted that its success will depend on how it is enforced, how swiftly it is enforced, and whether it targets the many misuses of AI and requires platforms to be more transparent with regards to their algorithms.
What more should be done to combat potentially harmful social media and AI content?
We would make a few suggestions:
● Digital watermarks could be required for AI-generated content, upon being uploaded to a platform.
● Age assurance could be mandated for AI-driven platforms, and in doing so prevent exposure to harmful content.
● Secure verification & e-signature tools for content uploads should be adopted by the industry, in order to foster a culture of safety and traceability.
● Regulators could increase and better coordinate their work on contents, although this should not be at the expense of other much-needed areas of focus.
Ofcom and the National Security Online Information Team are central in enforcing content moderation standards and reducing online harms. For this reason, it is key that they are well resourced, able to enforce rules, and that they focus on key areas. We welcome the recent publication of the Draft Statement of Strategic Priorities for online safety which aims to ensure that this is the case.
Lawmakers should also consider whether the current remit of these two organisations are sufficient to enable them to effectively supervise and audit AI algorithms.
We believe that accountability should be shared amongst the various members of the online industry, as well as those who create and upload content. Platforms should be responsible for algorithmic design and implementation. Content creators should be responsible for the integrity of the material they publish and share. Third-party developers should be responsible for the ethical development of AI tools, including the provision of adequate safeguards.
18 December 2024