AI’s Improved Intelligence Owes Much to Smarter Human Trainers!

Artificial intelligence (AI) models like ChatGPT and its rival Cohere have evolved from their early stages, where large teams of low-cost workers were employed to train the models on simple tasks, such as distinguishing basic facts, like identifying whether an image was of a car or a carrot.

As AI technologies have become more sophisticated, so too has the need for a new kind of human trainer—those with specialized knowledge across various disciplines.

This shift reflects the rapidly growing demand for AI trainers with advanced degrees and expertise in highly specialized fields.

One company at the forefront of this shift is Invisible Tech, which has positioned itself as a leading AI training partner.

Comment
byu/Tyler_Zoro from discussion
inaiwars

Founded in 2015 as a workflow automation company, Invisible Tech has become a key player in AI. Today, the company employs over 5,000 trainers across more than 100 countries, with many holding PhDs, Master’s degrees, and other specialized qualifications.

Invisible Tech has become an essential partner to AI companies such as Cohere, AI21, and Microsoft, helping train their AI models to reduce errors, known as hallucinations.

AI hallucinations occur when AI systems generate incorrect or misleading information, a problem that has plagued even the most advanced AI models.

Comment
byu/Tyler_Zoro from discussion
inaiwars

For example, in 2023, a Google chatbot provided inaccurate information about which satellite first took pictures of a planet outside Earth’s solar system.

Invisible Tech pays its trainers competitively, with rates ranging from $40 to $50 per hour, depending on the complexity of the work and the trainer’s location.

Trainers can earn up to $200 per hour in some high-expertise fields, such as quantum physics.

Comment
byu/Tyler_Zoro from discussion
inaiwars

This growing demand for AI trainers is creating new job opportunities for individuals with specialized knowledge in areas such as medicine, finance, and science, even for those who may not have coding skills.

The company’s journey in the AI space began in 2022 when OpenAI, ahead of the public launch of ChatGPT, sought its expertise to address hallucinations in early versions of the model.

Invisible Tech provided reinforcement learning with human feedback to improve the AI’s performance and reduce hallucinations.

Today, Invisible Tech collaborates with OpenAI and other major AI firms. Zhang of Cohere mentioned that Invisible’s trainers have helped their AI model improve at tasks such as retrieving relevant information from large data sets.

Comment
byu/Tyler_Zoro from discussion
inaiwars

This process involves hiring trainers with expertise in specific subjects, such as Swedish history or financial modeling, to teach AI models how to behave in specialized environments.

While Invisible Tech has become a key player in this space, it faces competition from other firms like Scale AI, which has also ventured into providing AI training services.

Comment
byu/Tyler_Zoro from discussion
inaiwars

Scale AI, last valued at $14 billion, counts OpenAI as one of its customers. Despite the competition, Invisible Tech has carved out a profitable niche in the industry, having been profitable since 2021.

The evolution of AI training has created new opportunities for individuals from a variety of academic backgrounds. Previously, AI training involved data-labeling work that required less qualification and paid significantly lower wages, sometimes as little as $2 per hour.

However, as AI models become more complex, the demand for trainers with advanced degrees in a variety of fields continues to rise, leading to better compensation and more specialized roles.

Comment
byu/Tyler_Zoro from discussion
inaiwars

The increasing need for specialized human trainers highlights the ongoing challenge AI companies face in reducing hallucinations and ensuring that their models produce accurate, trustworthy outputs.

Armed with advanced degrees and subject-specific knowledge, human trainers have become indispensable to this effort, working alongside AI developers to refine models and minimize errors.

The growing reliance on specialized human trainers underscores a critical reality in the AI industry: while AI models are becoming more powerful and capable, their success is still heavily dependent on the expertise and input of highly qualified human trainers.

For more news and insights, visit AI News on our website.