Kevin Korte is President of Univention North America, making sure you stay in control of your data, your company and your future.
With AI increasingly taking over menial or repetitive tasks at work, most office workers are eager to try it for themselves. But the vast selection of up-and-coming AI tools makes it daunting to evaluate this veritable app zoo and select the right tools. Bring Your Own AI (BYOAI) is an easy fix—at first glance. That’s because we have been there before: The experiences with Bring Your Own Device and Shadow IT have shown that this freewheeling strategy has some serious downsides.
Easy: Why Companies Might Want To Encourage It
It is indisputable that AI scheduling and note-taking tools have already taken over a significant portion of the secretarial duties imposed on knowledge workers. Large language models do some things really well. They shine in summarizing meetings, optimizing agendas and creating a list of follow-up tasks. Neither organizations nor employees want to miss out on the returns of those tools. BYOAI, in short, is a quick and easy way to achieve those types of productivity gains.
Moreover, getting on board with new AI tools comes at a relatively low cost to the company. End-user-focused tools are often cheaper than enterprise management features, and in many scenarios, the employee pays for the apps or tools. Companies will also save money on the initial selection process.
When end users pick their AI tools, they gravitate toward the most convenient ones, usually taking ease of use into account. By simply allowing this wisdom of the crowd approach to play out, IT departments can quickly identify and select the tool with the highest satisfaction rating.
Finally, when users use the same AI for their personal productivity tasks and for work, they self-train to optimize their AI usage. They become familiar with the interface and optimization strategies without the company having to spend significant amounts on training.
What’s more, there is less tool confusion as each team member has one tool for their specific tasks. As a result, employees will continuously hone their skills when using the various features of their chosen tool both at work and during personal projects.
Not So Easy: The Downsides Of Bringing Your Own AI
However, it’s just as important to weigh the potential risks of BYOAI. Like shadow IT, which refers to technology not officially sanctioned by the IT department, finding the right balance will be key to its success.
One of the most significant concerns with BYOAI is cybersecurity. With corporate data potentially scattered across multiple AI companies, it’s possible that the IT and cybersecurity teams might not even know which company currently holds your data. This means that a data breach at any AI company could have serious implications for your organization.
It’s, therefore, essential for the IT department to operate under the assumption that no data breach will compromise enough data to threaten the company’s survival, or they will need to monitor hundreds of services and resellers for breach notifications.
Another challenge with BYOAI is that it places the responsibility for privacy and data protection regulations on the individual employee. This means that employees, not the IT and legal departments, will need to spend time analyzing terms and conditions and privacy statements to determine whether or not company data, especially customer data, can be stored with a service provider.
BYOAI assumes, quite optimistically, that each individual employee will perform the same due diligence and keep on top of the monitoring whenever the company modifies the terms, the law changes or the courts reinterpret regulations. They might even be liable for the potential damages if they get it wrong, which places a significant burden on team members.
Lastly, AI systems are rapidly evolving. At the same time, a significant amount of media time focuses only on a small subset of large language models. No wonder that for the casual user, the leading models seem to provide the best results across every industry and application.
In reality, it’s not that simple. The emerging segment of industry-specific small language models and agents shows much more promise in their respective niches. Employees, therefore, might train themselves using inappropriate or outright wrong tools and achieve substandard results.
How To Make A Strategic Decision On Tolerating AI Tools
Shadow IT and Bring Your Own Device have significantly challenged the IT department. As it stands, the power of large language models and GenAI can substantially worsen the situation. At the same time, Bring Your Own AI might offer many companies and employees a quick way to increase productivity and profit from emerging trends.
Every company now faces a dilemma: Evaluate all the novel options for too long, and you risk losing out harvesting low-hanging fruit. Jumping on the AI bandwagon with too much haste, on the other hand, can conjure up a host of substantial problems. Making a conscious, strategic decision on tolerating, embracing or prohibiting BYOAI is critical to getting the most out of it.
To get started, find areas where you can quickly begin controlled tests. For example, your sales staff might spend much time researching prospects, and AI can easily take over that task. A customer services bot to answer emails around internal IT questions might give you another quick and easy win where you can see whether BYOAI or traditional approaches are easier to handle.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
