Many AI models like OpenAI’s ChatGPT can perform research for users. But OpenAI thinks that’s not good enough. In fact, OpenAI believes that by 2028, it will have developed a “legitimate AI researcher.”
OpenAI developing AI researcher
So as not to confuse you, by “AI researcher,” OpenAI is referring to an AI that can perform research. Not a human researcher that does research on AI. According to the company’s CEO, Sam Altman, OpenAI’s deep learning systems are advancing at such a rapid pace that the company is tracking towards achieving an intern-level research assistant by 2026. Then by 2028, it would have a “legitimate AI researcher.”
If that is true, it would be impressive. Like we said, AI models like ChatGPT have the ability to perform deep research for users. This means that instead of a cursory search, it dives deeper into the web to find information on the things you want. However, does it perform as well as a human? That’s debatable.
AI research versus human research
There’s no doubt that Deep Research will save humans a lot of time. Imagine you’re trying to write a thesis for your final semester at university. Normally, you’d have to go to the library and search for books related to your subject. Then within those books, you’d have to find information that’s relevant. You could do it online, but it’s also the same thing. You’d have to sift through dozens, if not hundreds of websites, to find information to support your paper.
With Deep Research, all of that information can be found and collated in minutes. This frees you up time on researching and more time writing. However, speed is the only thing it excels at.
When we do our own research, we learn things along the way. We employ critical thinking and we understand the context of the information. AI, however, does not. It’s great at summarizing information, but because it lacks the nuances of human thinking, the information it finds might not be as good as a human researcher doing the same thing.
Right now, tools like Deep Research are more like a jumping off point. It helps you find information and sources quickly, but then you’ll have to synthesize that information yourself. However, if OpenAI is capable of achieving human-level research in the next few years, it will be a whole different ball game.
The post OpenAI Aims to Build a Legitimate AI Researcher by 2028 appeared first on Android Headlines.