Experts On Why Comms Professionals Simply Should Not Be Using AI Model Deepseek

(MENAFN- PRovoke) In an era where artificial intelligence is rapidly transforming the communications landscape, the recent emergence of DeepSeek-R1-a generative AI model from a Chinese startup-serves as a stark reminder of the potential pitfalls associated with these advanced technologies.

According to a report by Enkrypt AI, DeepSeek-R1 is 11 times more likely to generate harmful content as compared to its established competitors, raising significant concerns about bias, toxicity, and security vulnerabilities.

As organizations increasingly turn to AI to enhance their messaging and streamline operations, communicators must tread carefully, weighing the benefits of innovation against the ethical implications of deploying models that may perpetuate harm.

Currently, generative AI adoption has approached an all-time high among communicators with a recent report by Muck Rack citing three out of four professionals using the technology. This is nearly three times the number from March 2023. 73% of Gen Z use the tool in daily tasks as well while more than half of PR pros say they use a paid version of AI services.

As such, the use cases for generative AI in communications are extremely well understood, according to Chad Latz, chief innovation officer at Burson. He added that models have been added as features into much of the commstech that we are using, or as independent utilities inside the enterprise.

"These use cases and capabilities are also true for DeepSeek. PR pros immediately grasp the ease with which they can quickly create content at scale, leverage answer engines to conduct research, utilize models to analyze complex data and track information, power creative ideation, build editorial calendars, run crisis scenarios, and more," he said.

Latz added that DeepSeek, like many of the recent models is boasting superior advanced logic capabilities. "While we may not see broad adoption by large global companies of DeepSeek models across the enterprise for reasons that are now being expressed. Communications professionals in the domestic Chinese market will now take full advantage of the latest innovation."

Saying that, Latz noted that when it comes to technology, what we see in the marketing and communications industry broadly, is a preoccupation with efficiency gains from generative AI. This will also be the case with DeepSeek.

"The opportunity however, as we see it, is not solely in using gen AI to execute and expedite human tasks, but rather in creating our innovative solutions using cognitive AI to predict how humans will think and behave, something far more valuable to companies and brands attempting to manage risk in a volatile social, political, and business environment," he said.

Saying that, industry experts are wary of how quickly it will, or rather, should, pick up among communicators. According to David Mayo, managing partner at NADA Creatve Intelligence, governments around the world need to beware of rising Chinese covert influence while consumers need to be aware of the quality of what they get.

“For example, if you ask DeepSeek about Taiwan and Tiananmen Square it wants to talk about something else, so the AI is not freeware as it seems to come under the Core Socialist Values of the National Cybersecurity Standards Committee,” he said.

True enough, just this week, Taiwan's Ministry of Digital Affairs (MODA) banned its public sector workers from using DeepSeek due to national information security concerns. The ministry noted that DeepSeek's operations involve cross-border transmission, information leakage, and other information security concerns.

As a result, experts such as Aaron Kwittken, founder, and CEO of PRophet, which has built an AI platform for the PR industry, blatantly say that comms professionals simply should not be utalising DeepSeek for their work.“ It's a Chinese-owned entity that is already facing self-censorship issues. If you were concerned about data privacy with US-based generative artificial intelligence (GAI) companies, then you should be terrified to use DeepSeek for anything serious or material. I would strongly discourage people from being insurgent and use it at or for work,” he said.

“GAI is yet just one component of comms tech and the rise of communications engineering. If you want to use GAI, then only use secure platforms that are purpose-built, secure, and customized for our workflows,” he said.

He added that as an industry, we will only continue to see more and more GAI alternatives to ChatGPT like DeepSeek popping up.“Start-ups like DeepSeek are actually likely to drive more interest and demand and drive improvements in GAI by competing based on performance, not just PR peacocking,” he said.

He added that the biggest takeaway is that GAI is and will become a commodity use case for AI.“To date, its been expensive to implement at scale so more competition should bring costs down and drive adoption up. It's also not great, but getting better. Humans will always be better, for now,” he said.

MENAFN05022025000219011063ID1109169576