AI training is booming — and LinkedIn wants a piece of the pie.
The careers networking site is in the early stages of launching an “AI labor marketplace” where people can make up to $150 an hour training AI chatbots to get better at everything from coding to nursing to finance, LinkedIn confirmed to Business Insider.
A spokesperson for LinkedIn, which is owned by Microsoft, said AI training is one of the fastest-growing jobs in the US right now and that it’s doing early testing.
AI trainers are humans who help improve chatbots by rating their answers and testing their limits. It’s a new type of gig work spurred by the AI boom and has led to the creation of several rapidly-growing AI training startups that serve clients like Anthropic.
LinkedIn has over a dozen public listings asking for AI trainers.
Someone with expertise in Excel and finance can make up to $100 an hour, while a nurse can make similar rates. The highest-paid position, for a senior software engineer AI trainer, pays up to $150 an hour. Other roles include a Germanic and Nordic Linguists trainer, which pays up to $100, and someone who “red teams” — or tests — AI systems for $40-$50 an hour.
LinkedIn has also rolled out a feature that lets people receive notifications whenever an AI training opportunity pops up.
The move puts LinkedIn in direct competition with a host of fast-growing AI training startups that match frontier AI labs like OpenAI with human talent to improve their models.
Mercor quintupled its valuation in less than a year to $10 billion. Another AI training startup, Surge AI, which owns the human-expert marketplace Data Annotation, is valued at $24 billion, Forbes reported.
The sector’s breakneck growth — and vast armies of contributors — have also contributed to serious cybersecurity issues.
Scale AI, for example, left confidential contractor and client information wide open across hundreds of Google Docs last year, locking them down after Business Insider revealed the practice. Mercor was recently hit by a serious data breach that compromised its contractors’ data and led to five class-action lawsuits in a single week.
Read the full article here
