The CEO of $2 billion AI training startup, Invisible Technologies, says humans will continue to play a crucial role in training artificial intelligence for decades, despite industry talk of synthetic data replacing human input.
Humans Remain Essential As AI Struggles With Complex Tasks
Last week’s episode of the “20VC” podcast, Matt Fitzpatrick said one of the biggest misconceptions in AI training is that human feedback will soon become unnecessary.
“When I first started this job, the main push back I always got was that synthetic data will take over and you just will not need human feedback two to three years from now,” Fitzpatrick said.
He added, “From first principles, that actually doesn’t make very much sense.”
Synthetic Data Can’t Replace Human Expertise
Synthetic data is artificially generated and often used when real-world data is scarce or restricted by privacy concerns.
Humans provide feedback by labeling, ranking, and correcting AI outputs, teaching models subtle skills such as empathy, humor, or context-specific reasoning.
Fitzpatrick noted that AI models still struggle with complex tasks that require language, cultural, and legal understanding.
“On the GenAI side, you are going to need humans in the loop for decades to come,” he said.
Invisible Technologies, valued at $2 billion, raised $100 million in September and competes with other data labeling firms such as Scale AI and Surge AI, which collectively employ millions of human contractors.
CEOs of similar startups have echoed Fitzpatrick’s view, emphasizing that high-quality, specialized human input remains essential even as AI models improve.
See Also: Trump’s AI Order Under Fire— Amy Klobuchar Says It’s ‘Likely Illegal’ While Bernie Sanders Calls It ‘Extremely Dangerous’
AI Hype, Job Risks And Ethical Concerns Shake Industry
Last month, Alphabet Inc. (NASDAQ:GOOGL) (NASDAQ:GOOG) subsidiary Google DeepMind CEO Demis Hassabis warned that many AI startups were massively overvalued, raising billions before fully launching products, and suggested a market correction could be coming.
He said AI was overhyped short-term but underappreciated long-term.
Meanwhile, AI advanced rapidly, and Geoffrey Hinton cautioned that it could replace millions of jobs by 2026, including call center and complex engineering roles, while raising concerns about AI’s ability to deceive.
Sen. Bernie Sanders (I-Vt.) also warned that AI and robotics could be dangerous if they mainly benefited big tech, urging that the technology be developed to improve human life rather than enrich the wealthiest or erode democracy and privacy.
Read Next:
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.
Image via Shutterstock
