Manipulated images and videos that appear "completely real" will be available to everyday people for "half a year to a year," Deep False Pioneer Hao Li told CNBC on Friday.
"It's still very easy, you can tell most of the deep subject with the naked eye," Li, a computer science professor at the University of Southern California, said of "Power Lunch."
"But there are also examples that are really, really compelling, "Li said, adding those that require" enough effort "to create.
"Deepfake" refers to the process of computers and machine learning software to manipulate videos or digital representations to make them appear real, even if they are not.
However, the rise of this technology has raised concerns about how these creations can cause confusion and disseminate information, especially in the context of global politics. Online disinformation through targeted social media campaigns and apps such as WhatsApp has already released choices worldwide.
Li's appearance on CNBC follows an appearance earlier this week at a Massachusetts Institute of Technology conference, where he said he thought perfect in-depth parties would arrive at "two to three years."
In an email to CNBC after asking for clarification, Li said the latest developments, especially the emergence of the hugely popular Chinese app Zao and the growing research focus, which have led him to "calibrate" his timeline.
"In some ways, we already know how to do it," Li wrote in an email, adding that it is just "a matter of training with more data and implementing it." [1
"Soon it will come to the point where there is no way we can actually discover [deepfakes] anymore, so we have to look at other types of solutions," Li said of "Power Lunch."
That's why research by academics is important, Li said, noting her work on discovering in-depth fake with Hany Farid, a professor at the University of California at Berkeley.  "If you want to be able to detect deepfakes, you also have to see what the limits are," Li said. "If you need to build AI frameworks that can detect things that are extremely real, they need to be trained using this kind of technology, so in some ways it's impossible to detect them if you don't know how they work."
For Li, the problem with deepfakes is not the existence of the technology that can create them.
He said that deepfake technology provides many benefits to the fashion and entertainment industry, for example. It can also improve the effectiveness of video conferencing, Li said.
"The real question is how can we detect videos where the intent is something that is used to deceive people or something that has a harmful consequence," he said.