As an industry-first approach to commercially safe AI, Shutterstock TRUST serves as a guide for brands and agencies seeking a framework for responsible AI
Shutterstock, Inc., a leading global creative platform connecting brands and businesses with high-quality content, launched TRUST, its official framework based on the long-standing Shutterstock’s values and efforts to use AI responsibly in its products and workflows. TRUST, an acronym for Training, Royalties, Uplift, Safeguards and Transparency, represents the most important goals that Shutterstock has actively pursued over the past two decades. In an unregulated and rapidly evolving industry, this framework can also serve as a practical checklist for companies experimenting with AI technologies and underscores Shutterstock’s trusted, vendor-centric principles.
TRUST’s goals include:
Training using properly licensed data
Fair pricing for creative
Boosting representation that reflects the world
Safeguards to control content and protect customers
Transparency by default
“Maintaining the trust of our vendors and customers is our focus, and our TRUST framework reinforces our commitment to achieving this goal,” said Dr. Alessandra Sala, senior director of AI and data science at Shutterstock. “We have always Properly licensed data used for training purposes, new revenue streams for our vendors, and targeted searches for highly motivated creatives and underrepresented groups. TRUST combines these principles into a clear and straightforward approach that other companies in the industry can follow as they chart their own path to stewardship “Use AI.”
Training
As a global leader in content licensing, Shutterstock respects creators’ intellectual property rights. This includes measures to ensure that our AI models are only trained using properly licensed content and that our AI partners also meet this standard. Shutterstock takes targeted steps to ensure that it has the rights to license the content on its platform and that the content meets its standards.
In addition, providers have the option to opt out of data contracts to respect their autonomy and preferences. However, so far only around 1 percent of all Shutterstock providers have decided to do so.
Shutterstock
‘s Contributor Fund is a testament to our commitment to supporting creators through compensation. To date, the Shutterstock Contributor Fund has paid out hundreds of thousands of creators, and is expected to compensate millions more artists for using their content to train our generative technology.
Elevating
Our mission has always included promoting inclusive content and investing in collaboration with underrepresented content creators to achieve the broadest possible range of representation. Shutterstock teams are committed to increasing representation across genders, ethnicities and orientations and ensuring a fair creative platform through initiatives such as Create Fund and partnerships with UN AI for Good and World Ethical Data Foundation. This includes improving and expanding stock content created by humans so that our AI capabilities enable advanced use of such content and support, rather than replace, the unique talent of artists.
Safeguards
Shutterstock has pioneered pioneering safeguards that ensure a peace-of-mind experience for its customers and their audiences through tailored safeguards and content controls, as well as an ethical approach, legal integrity and brand safety.
Shutterstock’s layered content controls help prevent the creation of content that contains sexual or offensive themes or violence. These safeguards also include protection against liability for any licenses of AI-generated content related to AI training. Shutterstock Enterprise users can receive the same protection for AI-generated content as traditional stock content, while special cases are covered by an extended content warranty.
In a survey of Shutterstock Enterprise customers, more than 60 percent of respondents said ensuring secure commercial use was the most important criterion for selecting an AI platform. Shutterstock is the only company to offer enterprise users the option of human review of their licensed content for compensation.
Transparency
Shutterstock believes in transparency around AI practices and content provenance. The company is committed to clearly distinguishing between AI-generated and human-generated content on the Shutterstock platform to protect users from deepfakes, misinformation, and other forms of abuse. Shutterstock does not sell AI-generated content with unknown or unclear origins. Providers cannot upload AI-generated content to Shutterstock’s main library. As a supporter and member of the Content Authenticity Initiative, the company also labels all AI-generated content and metadata according to C2PA/IPTC industry standards to further protect its users with verifiable and tamper-proof information on all content types. This ensures that each asset is certified with secure metadata about its creation, authorization and change history.
Shutterstock’s ethical approach to AI has made the company a sought-after collaboration partner with industry leaders such as NVIDIA, Meta, OpenAI, LG and others to provide fundamental generative AI tools and standards for 3D, image and text creators develop. Shutterstock’s high-quality content library, enriched with rich metadata, leads the industry in size, diversity, and annotation, making it unrivaled for training AI skills.
SOURCE: PRNewswire