Invisible Tech Becomes A Leader In AI Training, Recruits Thousands Of Global Experts

AI is now supported by smart human training (photo; dock. pexels)

Jakarta - Along with the rapid development of artificial intelligence (AI) technology, the need for human trainers who have deep knowledge in the special field is increasing. In the past, AI model training such as ChatGPT and Cohere only needed cheap labor to help distinguish basic facts, such as car images or carrots. However, the latest more sophisticated AI model requires trainers with specific knowledge in various fields, ranging from history to medical science, including experts with PhD degrees.

Invisible Tech, one of the companies now leading the provision of human trainers for AI models, has employed more than 5,000 trainers from more than 100 countries around the world. The company is working with major companies such as OpenAI, Microsoft, and AI21 to train AI models to be more accurate and reduce errors known as "AI halusinations". AI halusination is a condition in which AI models produce misinaccurate or inaccurate information, which can reduce user confidence in this technology.

Ivan Zhang, co-founder of Cohere, explained that in the past they could recruit students to train AI models in general. However, now their AI models require training from professional experts such as doctors and financial analysts to handle complex scenarios in the medical or financial fields.

"Invisible Tech is the main AI training partner for many major companies such as AI21 and Microsoft," said Francis Pedraza, founder of Invisible Tech, quoted by VOI from Reuters. "We have 5,000 coaches specializing in various fields, including PhD and master's degree holders."

Invisible Tech is not only working with OpenAI, but also with many other AI companies. Invisible Tech claims that these companies use human trainers to reduce the burden of AI errors that often appear when AI cannot distinguish between fact and fiction.

High Income For AI Coaches

Invisible Tech offers varying payments depending on the location and level of worker expertise. Workers in some companies can earn up to 40 US dollars per hour, and in some cases even up to 200 US dollars per hour for very specific subjects, such as quantum physics. With the increasingly sophisticated AI models, the need for human trainers with high specializations also continues to increase.

Invisible Tech was originally founded in 2015 as a company that automates workflows for clients such as DoorDash, but their roles changed drastically when OpenAI, a research company that later released ChatGPT, asked for their help in 2022. At that time, OpenAI faced major problems: the initial ChatGPT model often produced unreliable or hallucinatory information. Invisible Tech then helped train this model to be more accurate.

Apart from Invisible Tech, other companies such as Scale AI are also competing in providing human trainers for AI companies. The AI scale, valued at around $14 billion, offers similar services and is also one of OpenAI's training providers. Even so, Invisible Tech has been one of the main players in the market with stable profits since 2021, despite only receiving funding of 8 million US dollars (Rp120.9 billion).

According to Pedraza, Invisible Tech remains the company's top choice AI because of their ability to provide high-quality training without burdening these companies with additional management costs.

The Future Of AI Training

With the increasingly complex needs of AI companies, demand for experts who can train AI continues to increase. In fact, many new companies have emerged to provide coaching personnel who can help create high-quality data for AI laboratories. Zhang, co-founder of Cohere, revealed that many new companies in this field continue to emerge and create job opportunities for those with specialized expertise.

"My Inbox was flooded with emails from new companies looking to offer their services to train AI," Zhang said.

With increasing demand, the future of AI training seems to be met with beneficial job opportunities for experts in various fields. The combination of advanced technology and smart humans is key to producing more reliable and accurate AI models in the future.