Key Takeaways:
- Artificial intelligence won’t be training AI anytime soon, and human feedback will still be needed for decades to come.
- The data labeling industry will continue to require human effort, with a shift towards hiring highly specialized experts in math and science.
- Synthetic data, which is artificially created, is used for training AI or machine learning models, but it has its limitations and cannot replace human feedback entirely.
- The industry will continue to require human contractors to help teach AI models various skills, including math, science, coding, and characteristics such as humor and empathy.
Introduction to the Topic
The CEO of Invisible Technologies, Matt Fitzpatrick, recently appeared on the "20VC" podcast, where he discussed the future of the AI training industry. Fitzpatrick emphasized that one of the biggest misconceptions in the industry is that humans won’t be needed in a few years. He stated, "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. From first principles, that actually doesn’t make very much sense." This statement highlights the importance of human feedback in the AI training process, and Fitzpatrick’s perspective on the industry’s future.
The Limitations of Synthetic Data
Synthetic data refers to data that is artificially created, and it is used for training AI or machine learning models, mostly where real data is scarce or can’t be used because of privacy concerns. However, Fitzpatrick notes that there are too many kinds of tasks for AI to accomplish in the world, and it would take a long time to do them accurately with language and cultural context in mind. For example, the legal industry contains vast amounts of nonpublic information, which makes it challenging to rely solely on synthetic data. As Fitzpatrick said, "On the GenAI side, you are going to need humans in the loop for decades to come. And I think that is something that most people are starting to realize."
The Importance of Human Feedback
Human feedback is essential in the AI training process, as it allows real people to filter, rank, and train AI responses. Fitzpatrick’s company, Invisible Technologies, competes with data labeling companies such as Scale AI and Surge AI, which have raised billions in the past year as tech giants race to secure the data needed to train their AI models. These startups hire millions of human contractors, who help teach the models math, science, coding, and characteristics such as humor and empathy. As Fitzpatrick noted, "You are going to need humans in the loop for decades to come," emphasizing the ongoing need for human feedback in the industry.
The Shift Towards Specialized Experts
The data annotation industry is shifting from requiring generalists to highly specialized experts, including in math and science. Garrett Lord, the CEO of Handshake, a job platform that pivoted into AI training last year, said, "Now these models have kind of sucked up the entirety of the entire corpus of the internet and every book and video. They’ve gotten good enough where, like, generalists are no longer needed." This shift towards specialized experts highlights the increasing complexity of AI models and the need for highly skilled workers to train them.
Conclusion
In conclusion, the AI training industry will continue to require human effort, with a focus on hiring highly specialized experts in math and science. Synthetic data has its limitations, and human feedback is essential for training AI models accurately. As Fitzpatrick said, "From first principles, that actually doesn’t make very much sense" to rely solely on synthetic data. The industry will continue to evolve, with a growing need for human contractors to help teach AI models various skills. As the CEO of Mercor, Brendan Foody, noted, "having phenomenal people that you treat incredibly well" is crucial for the success of the business. The future of the AI training industry will depend on the ability to balance the use of synthetic data with human feedback, and to adapt to the changing needs of the industry.
https://www.businessinsider.com/ai-training-ceo-artificial-data-humans-matt-fitzpatrick-invisible-technologies-2026-1