One of the irresistible things about generative AI is how it makes us feel smart. We can ask ChatGPT or Copilot a clever question and instantly get a satisfying in-depth response that makes us feel like we have all the answers. There’s a catch, though — large language models tend to make up things that sound viable but aren’t true; they aren’t always good at reasoning; and they are often outright biased. If we blindly put our trust in AI output, we risk looking dumb.

AI models still lack thinking skills, which means we humans have to demand more from our own analytical minds. To use GenAI effectively, we have to learn how to discern where it’s helpful, where it’s misleading, and what output to outright reject. But we also have an opportunity to share our critical thinking skills and shape AI models at the development stage.

A great deal of the success or failure of AI models depends on the data they are trained on. Large language models learn from huge datasets of sample texts and feedback from human experts. These systems are evolving fast, and they are in dire need of human intervention to improve the quality of output and eliminate bias, factual errors, logical inconsistencies, and unethical responses.

At Mindrift, we are serious about making sure that a model is trained on the best possible data. Mindrift AI Tutors (formerly known as AI Tutors by Toloka) are teams of industry professionals who apply their critical thinking skills and expertise to improve the quality of texts for training GenAI. As one of our editors, Feyaza, explains: “Fact-checking, source analysis and other critical thinking skills are crucial for properly training AI. These applications of critical thinking, especially when combined with an AI trainer's own expertise, make all the difference in ensuring accuracy and preventing bias. Yes, LLMs can offer amazing output. But human expertise and oversight – especially those critical thinking skills – are essential to guide AI capabilities in the right direction.”

Feyaza and other AI Tutors recently shared their insights with Forbes magazine regarding the crucial role of human data trainers in AI development. For a deeper dive, read the full story on critical thinking and GenAI: In The Age Of AI, Critical Thinking Is More Needed Than Ever.