Ed Newton-Rex, former head TikTok AI designer and executive at Stability AI, has launched a nonprofit called Fairly Trained. Fairly Trained offers a certification program, called L Certification, for AI companies that license their training data. Through this program, companies are required to demonstrate that their training data was properly licensed, in the public domain, offered under an open license, or owned by the company.
Already, nine companies including Bria AI and LifeScore Music have received certification. The nonprofit charges a fee ranging from $500 to $6,000 based on the company’s size. While OpenAI argues the need for unlicensed data to create generative AI services, Fairly Trained and its certified companies firmly disagree. With support from trade groups and companies like Universal Music Group, Fairly Trained is gaining momentum in challenging the standard approach of the AI industry. Although currently a one-man operation, Newton-Rex has plans to introduce additional certificates in the future, potentially addressing concerns regarding compensation.
Table of Contents
Former TikTok AI Designer Launches Fairly Trained Nonprofit
Fairly Trained: A Nonprofit for Ethical AI Training Data
Introduction to Fairly Trained
Fairly Trained is a non-profit organization founded by Ed Newton-Rex, a former head AI designer at TikTok and executive at Stability AI. The organization aims to address the ethical concerns surrounding AI training data and ensure the development of responsible and unbiased AI systems. By offering a certification program known as L Certification, Fairly Trained aims to promote transparency, accountability, and fairness in the AI industry.
Founder’s Background in AI Design
Ed Newton-Rex brings a wealth of expertise in AI design to Fairly Trained. Having worked as the head AI designer at TikTok and as an executive at Stability AI, Newton-Rex has a deep understanding of the challenges and opportunities in the AI industry. His experience in creating AI systems and working with large-scale platforms gives him unique insights into the potential biases and ethical issues that can arise in AI training data.
Mission and Goals of Fairly Trained
Fairly Trained’s mission is to prioritize the use of ethical training data in the development of AI systems. The organization aims to ensure that AI technologies are built on fair and unbiased foundations, thereby promoting trust and accountability in the industry. Fairly Trained also strives to advocate for industry-wide change, pushing for ethical standards to become the norm rather than the exception.
Certification Program: L Certification
L Certification is the flagship program offered by Fairly Trained. It provides a way for AI companies to demonstrate their commitment to ethical practices in training data. Companies seeking certification undergo a rigorous evaluation process to ensure compliance with Fairly Trained’s criteria for responsible data usage. By obtaining L Certification, companies can assure their stakeholders that they have taken concrete steps to mitigate biases and ensure fairness in their AI models.
Requirements for Certification
To receive L Certification, companies must meet certain requirements set by Fairly Trained. These requirements include proving that their training data was explicitly licensed, obtained from the public domain, offered under an open license, or already owned by the company itself. By adhering to these requirements, companies demonstrate their commitment to using data from reliable and ethical sources.
Companies Certified by Fairly Trained
Fairly Trained has already certified nine AI companies, including Bria AI and LifeScore Music. These certified companies have met the rigorous standards set by Fairly Trained and have demonstrated their dedication to ethical data usage. By obtaining certification, these companies gain recognition and credibility within the industry, distinguishing themselves as leaders in responsible AI development.
Certification Fees Based on Company Size
Fairly Trained charges a certification fee based on the size of the company seeking certification. This fee structure ensures that the cost of certification aligns with the resources and scale of the company in question. Small companies pay a fee of $500, medium-sized companies pay $2,500, and large companies pay $6,000. This tiered fee structure accommodates companies of varying sizes and ensures accessibility to the certification program.
Debate with OpenAI: Use of Unlicensed Data
OpenAI, a prominent player in the AI industry, holds a different perspective on the use of unlicensed data. OpenAI argues that it is impossible to create generative AI services without utilizing unlicensed data. However, Fairly Trained and the certified companies disagree, insisting on the importance of using explicitly licensed, public domain, or open-licensed data to ensure ethical practices. This ongoing debate highlights the varying viewpoints within the industry and the need for continued discussions on responsible AI development.
Support from Trade Groups and Companies
Fairly Trained has garnered support from trade groups and companies that share its vision for ethical AI development. Universal Music Group, among others, has expressed support for the organization’s efforts to promote responsible data usage. However, while support from these industry players is valuable, the movement to change the standard approach to AI development is still in its early stages. Fairly Trained continues to work towards gaining broader support and initiating industry-wide change.
Future Plans for Fairly Trained
As a one-man operation, Ed Newton-Rex envisions further expansion and growth for Fairly Trained. The organization aims to offer additional certificates in the future, potentially addressing issues like compensation in the AI industry. By continuously evolving and adapting to the changing landscape of AI development, Fairly Trained aims to play a pivotal role in shaping ethical practices and standards within the industry. The organization also seeks opportunities for collaboration and growth, recognizing the need for collective efforts to bring about meaningful change.
Introduction to Fairly Trained
Overview of Fairly Trained
Fairly Trained is a nonprofit organization founded by Ed Newton-Rex with the goal of addressing the ethical concerns surrounding AI training data. The organization’s primary focus is to promote responsible and unbiased AI systems by offering a certification program, known as L Certification, to AI companies.
Importance of Ethical AI Training Data
The use of ethical training data is crucial to ensure the fairness and reliability of AI systems. AI models heavily rely on training data to learn and make predictions, and the quality and nature of this data can significantly impact the outcomes of AI algorithms. Ethical training data helps mitigate biases, discrimination, and unfair practices, fostering trust and accountability in AI technology.
Context of AI Industry Standards
The AI industry is constantly evolving, with new technologies and applications being developed at a rapid pace. However, ethical standards and practices have lagged behind, leading to concerns regarding biased algorithms and the potential for discrimination. Fairly Trained aims to address these issues by raising awareness and establishing industry-wide guidelines for responsible AI development.
Need for Nonprofit Solutions
The establishment of Fairly Trained as a nonprofit organization highlights the urgency and necessity of addressing the ethical concerns in AI training data. Nonprofit entities can operate with a focus on promoting the public interest rather than pursuing profits, allowing them to advocate for ethical practices and drive positive change in the AI industry. Fairly Trained fills this crucial role, providing a platform for AI companies to demonstrate their commitment to responsible data usage.
Related site – TikTok’s Viral Challenges Keep Luring Young Kids to Their Deaths