A new Artificial Intelligence learning model, Deepseek, arrived on the global AI scene this week. Deepseek, a new AI model similar to Chat GPT or Google Gemini, was made by a company based in China.
In response to this, BYU-Idaho wrote a notice that appeared on canvas Tuesday, Jan. 28.

“There’s a company claiming credit for this, that they’ve created,” said David Ashby, BYUI professor and Director of Learning Innovation and Technology. “And they’re claiming that they’ve built the whole thing for $5 million, which is ridiculously low.”
The US AI market crashed shortly thereafter, with Nvidia stock dropping upwards of $600 billion on Monday, Jan. 27.
“They’re also saying that it's not their primary business model,” Ashby said. “Deepseek is a side project, just like Gmail was a side project of Google. So everyone’s highly skeptical of what they’re claiming right now.”
The reason for this chaos, according to Ashby, was the fact that Deepseek allegedly cost only $5 million to make.
Ashby said that OpenAI and other tech companies have spent hundreds of millions, and into the billions, to produce their large language models. Little is known about Deepseek, causing this disruption in the AI scene.
“In fact, what came out (on Jan. 29) is that they may have trained their model based off of training from OpenAI, which means they basically violated OpenAI’s terms of service,” Ashby said. “There’s some really interesting back and forth going on, and no one knows what to believe just yet.”
According to Ashby, Deepseek is mostly similar to other learning models, but there is some “anecdotal evidence” of Deepseek censoring information about China.
“The terms of service even say that if you violate anything…you’ll be subject to Chinese law and Chinese courts.” Ashby said.
Ashby also wrote the notice that appeared on Canvas Tuesday, Jan. 28. According to Ashby, it was out of caution about privacy and safety for BYUI students.
“A lot of students will use their BYUI email addresses and accounts to create accounts with other third party tools like Chat GPT, and that’s a common practice,” Ashby said. “With this being such a new company and no one knows anything about it, and the terms of service being highly questionable, we just thought it was good to just preemptively kind of teach everybody that this is a brand new tool and nobody knows much about their background.”

Ashby said that the risk may be small, but they do encourage students to exercise caution.
“To share any kind of information with them, even your email addresses, opens up certain amounts of risk. And most of it is very small, minimal risk, but it opens risk,” Ashby said. “So we thought, for this particular case, until we know more information, it would be good to go out there and say, ‘We’re not recommending, in fact, we’re encouraging you not to use your byui.edu email domain to create accounts with this particular company’.”
Ashby mentioned a talk given by Elder David A. Bednar at BYUI on Nov. 3, 2024, where Elder Bednar spoke about AI and moral agency.
“All students need to watch Elder Bednar’s talk on ‘Things as They Really Are 2.0,’ Ashby said. “Elder Bednar speaks very plainly and says to make sure that they’re acting themselves and not being acted upon. So don’t let AI do your homework for you (or) write things for you.”
According to Ashby, it is absolutely crucial that people learn to think and write for themselves, and shouldn’t use AI to do all the work for them.
“Lots of times we use writing to think, and thinking is going to be critical in this next age,” Ashby said. “Don’t let AI do the thinking for you, make sure you’re involved.”
For more about BYUI’s AI policy, visit the Gen AI webpage.
Featured image: Deepseek’s logo. Photo Credit: Freemalaysiatoday.com





0 Comments