WX trends: DeepSeek and the rise of shadow AI
Our workplace experience (WX) trends series looks at recent news articles, videos, social media posts, and thought leadership pieces on workplace experience. You’ll also hear from our experts on what’s trending.
In this Q&A with Appspace Chief Product Officer Stan Stephens, we take a look at DeepSeek – the open-source AI platform from a two-year-old Chinese startup that’s taken the tech world by storm. Its rapid ascent, even topping Apple’s US App Store charts, highlights the allure of readily available, powerful and less expensive AI solutions. But this accessibility also raises questions about security, misinformation, and the rise of shadow AI.
To understand the implications of this trend, we asked Stan to give us his take on DeepSeek’s disruptive arrival and its potential risks.
The arrival of DeepSeek is a big moment for AI. How do you see it affecting the workplace?
DeepSeek has certainly got everyone talking! AI platforms, in general, mark a significant shift for workplaces. AI capabilities are already changing how we work and what we do. It won’t be long before many repetitive tasks will be automated, reshaping many entry to mid-level positions. And the arrival of DeepSeek will likely accelerate AI development. It’ll boost competition in the space and that will all have a knock-on effect on how AI is integrated into the workplace.
We’ll get into the issues specific to DeepSeek in a bit, but I want to be clear that I think AI will transform jobs rather than take them away. Embracing AI tools at work – that’s smart tech helping you communicate better, connect more easily with colleagues, translate information into multiple languages, and do those time-consuming admin tasks, for example – means employees can move into roles that require more critical thinking, creativity, and strategic problem-solving. So, rather than replacing people, AI will be a powerful tool that enhances our work. And that’s exciting. I can’t stress enough the importance of AI literacy and I encourage organizations to focus on it.
As AI performs more tasks, skills like emotional intelligence and human-AI interaction will become increasingly valuable. So, the challenge for leaders is to create an environment where AI and human expertise work together. And then we’re into using AI to transform learning and upskilling. AI can look at someone’s progress and learning styles and tailor training to work for them. It can provide real-time feedback and even use predictive analytics to determine future skill requirements.
We’ve already integrated AI into Appspace and it’s doing a lot of the heavy lifting at work – it’ll learn your preferences, for example, and help you solve issues based on your preferences and work style.
While DeepSeek is generating excitement, it’s also raising concerns about shadow AI (employees’ unsanctioned use of AI apps and tools without the oversight of IT). Why?
The popularity of new shiny tools like DeepSeek is understandable. We’re talking about tech that promises increased productivity and efficiency. People are naturally drawn to such solutions. However, this enthusiasm can inadvertently lead to shadow AI, where employees use unapproved AI tools, bypassing IT oversight. And that’s where the real risk lies.
The ease of access and the perceived benefits can tempt employees to use unsanctioned tools without considering the potential consequences for the organization. This risk is amplified by the fact that AI models, like DeepSeek, can be “black boxes.” In other words, without knowing how they arrive at their outputs, it can be challenging to identify and mitigate potential biases or even malicious content.
What are the biggest concerns you have about shadow AI?
We spoke before about data privacy, and my position on that hasn’t really changed. Having the right policies and governance in place – and being transparent about that – is what’s important.
Beyond data privacy, the potential for culturally misaligned or even unsafe content generated by AI models is a serious concern. Trusting a model requires not just an evaluation of its benchmark performance, but also an understanding of its origins and training data.
Businesses need robust change management strategies – clear communication, hands-on training, and a culture that embraces AI as an enabler, not a threat, while also emphasizing the importance of security protocols.
How can Appspace help organizations address the shadow AI challenge related to tools like DeepSeek?
Appspace is critical in facilitating crucial communication, with employee comms features built for sharing information. Our platform provides a centralized communication hub where organizations can easily disseminate information about IT policies, security best practices, and the risks of shadow AI.
Through targeted messaging, interactive content, and engaging campaigns, Appspace can help organizations raise awareness about the importance of using approved tools and following established security protocols.
We can also facilitate two-way communication, allowing employees to ask questions and express their concerns about AI adoption. This helps create a transparent environment where employees understand the “why” behind security policies.
And when it comes to AI, we’ve built it right into the core of our platform, so it’s right there.
In a nutshell, what’s your advice to businesses preparing for this AI-driven future?
Whether you’re looking into DeepSeek, ChatGPT, Gemini, or any of the AI platforms, I’d advise people to go into it with their eyes open. Embrace AI early, invest in your people, focus on adaptation. The businesses that succeed will be the ones that see AI as an opportunity to boost human potential rather than replacing it.
The future of work isn’t about AI versus humans – it’s about AI and humans working better together. But it’s also about ensuring that this collaboration is secure and controlled. Communicate clearly about AI policies, educate employees about the risks of shadow AI, and provide them with approved tools and resources. Critically, consider the source and potential biases of the AI models you adopt.
A balanced approach, combining innovation with robust security practices, is essential.
