• The AI industry now requires human trainers with specialized knowledge, from historians to scientists, to improve AI models.
• Cohere, a $5 billion AI company, collaborates with Invisible Tech, which employs thousands of remote trainers to reduce AI errors.
• Invisible Tech, initially a service for companies like DoorDash, became a key AI training partner after working with OpenAI to improve ChatGPT.
• As AI advances, the demand for specialized human trainers is growing, shaping the future of AI development across various industries.

In the rapidly evolving world of artificial intelligence (AI), the role of human trainers has become increasingly crucial. The early years of AI development saw models like ChatGPT and its rival Cohere relying on vast teams of low-cost workers to help distinguish basic facts. However, the landscape has changed dramatically. Today, the fiercely competitive arena of AI demands a rapidly expanding network of human trainers with specialized knowledge, ranging from historians to scientists, some even holding doctorate degrees.

Cohere co-founder Ivan Zhang, speaking about the company's internal human trainers, said, A year ago, we could get away with hiring undergraduates, to just generally teach AI on how to improve. Now we have licensed physicians teaching the models how to behave in medical environments, or financial analysts or accountants. This shift in hiring practices underscores the increasing complexity and sophistication of AI models.

Cohere, valued at over $5 billion, collaborates with a startup called Invisible Tech for additional training. Invisible Tech, employing thousands of remote trainers, has emerged as a key partner for AI companies ranging from AI21 to Microsoft. Their mission is to train AI models to reduce errors, known in the AI world as hallucinations.

The Role of Invisible Tech in AI Training

Francis Pedraza, Invisible's founder, said, We have 5,000 people in over 100 countries around the world that are PhDs, Master's degree holders and knowledge work specialists. The pay for these trainers can be as high as $40 per hour, depending on the location of the worker and the complexity of the work. Some companies, like Outlier, pay up to $50 per hour, while Labelbox pays up to $200 per hour for high expertise subjects like quantum physics.

Invisible, founded in 2015, initially catered to companies like food delivery company DoorDash to digitize their delivery menu. However, their trajectory changed when OpenAI contacted them in the spring of 2022, ahead of the public launch of ChatGPT. OpenAI approached Invisible with a problem: the early version of ChatGPT was prone to hallucinations, generating false outputs. They needed an advanced AI training partner to provide reinforcement learning with human feedback.

This partnership marked a significant turning point for Invisible, which has since become a training partner for most GenAI companies, including Cohere, AI21, and Microsoft. OpenAI, which ignited the frenzy around GenAI, has a team of researchers called the Human Data Team that works with AI trainers to gather specialized data for training its models like ChatGPT.

The Future of AI Training

They run dozens of experiments to improve writing style and reduce hallucinations. Based on the AI companies' needs, Invisible hires workers with relevant degrees for those projects, reducing the burden of managing hundreds of trainers by the AI companies. The competition in this space is fierce. Scale AI, a private start-up valued at $14 billion, provides AI companies with sets of training data and has ventured into providing AI trainers.

Invisible, profitable since 2021, has raised only $8 million of primary capital. Pedraza said, We are 70% owned by the team, and only 30% owned by investors. As AI companies launch more advanced models, the demand for specialized trainers and across dozens of languages is on the rise, creating a well-paid niche where workers from a variety of subjects could become AI trainers without even knowing how to code.

This demand is leading to the creation of more companies offering similar services. Zhang said, My inbox is basically inundated with new firms that pop up here and there. I do see this as a new space where companies hire humans just to create data for AI labs like us.

The evolution of AI training reflects the broader trend of AI becoming more sophisticated and specialized. The role of human trainers in this process is critical, as they provide the nuanced understanding and expertise that AI models need to function effectively in complex environments. As AI continues to advance, the demand for specialized human trainers is likely to grow, shaping the future of AI development and its applications across various industries.