Why employees smuggle AI into work

“It's easier to get forgiveness than to get permission,” says John, a software engineer at Financial Services Technology Company. “Try it and if you get into trouble later, clear it.”

He is one of many people who use their own AI tools in the workplace without the permission of the IT department (and therefore not using John's full name).

According to a study by Software AG, half of all knowledge workers use personal AI tools.

This study defines knowledge workers as “people who work primarily at desks and computers.”

While IT teams don't provide AI tools, others say they want to choose the tools themselves.

John's Company offers Github Copilot for AI-supported software development, but prefers cursors.

“It's mainly a glorious autocomplete, but it's very good,” he says. “It completes 15 lines at a time, then you look around it and say, 'Yes, that's what I would have typed out.” It will free you.

His misuse is easier than putting a lengthy approval process at risk, rather than violating the policy, he says. “I'm too lazy and I have enough pay to chase the costs,” he adds.

John recommends that businesses remain flexible in their choice of AI tools. “I've been telling people at work not to renew their team licenses one year at a time, as the entire landscape will change three months later,” he says. “Everyone wants to do something different and you'll feel trapped in the sinking costs.”

The recent release of Deepseek, an AI model freely available from China, could expand AI options.

Peter (not his real name) is a product manager for a data storage company and provides Google Gemini AI chatbots to people.

External AI tools are forbidden, but Peter uses ChatGpt via the search tool Kagi. He feels that the biggest advantage of AI comes from challenging his thinking when asking the chatbot to respond to his plans from the perspective of various clients.

“AI isn't giving you the answers as much as offering sparring partners,” he says. “As a product manager, you have a lot of responsibility and there are not many good outlets to openly discuss your strategy. These tools make it possible with free and unlimited capabilities.”

The version of ChatGpt he uses (4o) can analyze the video. “You can get a summary of your competitors' videos and have a whole conversation [with the AI tool] About the points of the video and how they overlap with your own product. ”

In a 10-minute ChatGpt conversation, he can see the material that takes him 2-3 hours to watch the video.

He estimates that his productivity increases will amount to a company getting a third of the additional people working for free.

He doesn't know why the company banned external AI. “I think it's about control,” he says. “Companies want to have a say in the tools their employees use. It's a new frontier and just want to be conservative.”

The use of unauthorized AI applications is sometimes referred to as “Shadow AI.” This is a more specific version of “Shadow It.” This is if someone uses the software or has not approved by the IT department.

Harmonic security helps identify Shadow AI and prevents corporate data from being inappropriately entered into AI tools.

It tracks over 10,000 AI apps, of which over 5,000 are used.

These include custom versions of CHATGPT and business software that adds AI features such as the communications tool Slack.

No matter how popular it is, Shadow Ai comes with risk.

Modern AI tools are built by digesting huge amounts of information in a process called training.

Approximately 30% of application harmonic security uses trains using information entered by users.

This means that user information will become part of the AI ​​tool and may be output to other users in the future.

While companies may be concerned that AI tools answers will make trade secrets public, Alastair Paterson, CEO and co-founder of Harmonic Security, thinks that is unlikely. “It's pretty difficult to straighten data from these [AI tools]” he says.

However, businesses are concerned that data is stored in AI services that are uncontrollable, unrecognized and vulnerable to data breaches.

Companies can be extremely useful, especially for younger workers, making it difficult to combat the use of AI tools.

“[AI] “We're committed to providing a great opportunity to help you,” said Simon Haighton-Williams, CEO of the UK-based software services group Adapavist Group.

“It cannot be completely replaced [experience]But just as having a good encyclopedia or calculator allows you to do things you couldn't do without those tools, it's a good leg. ”

What does he say to the companies that discovered they have Shadow Ai use?

“Welcome to the club, I think everyone is probably doing that. Be patient, understand what people are using and why, and understand how you can accept and manage it rather than demand it. [adopted AI]. ”

Trimble provides software and hardware to manage data about the built environment. To ensure employees can use AI safely, the company created the Trimble Assistant. This is an internal AI tool based on the same AI model used in ChatGPT.

Employees can consult with Trimble Assistant on a wide range of applications, including product development, customer support, and market research. For software developers, the company offers Github Copilot.

Karoliina Torttila is the director of AI at Trimble. “I encourage everyone to explore all sorts of tools in their lives, but I recognize that their professional life is a separate space and there are some protective measures and considerations there,” she says.

The company encourages employees to explore new AI models and applications online.

“This brings us to the skills we are all forced to develop. We need to be able to understand what sensitive data is,” she says.

“There must be a place where you can't put your medical information and you should be able to make such a type of judgement call [for work data, too]. ”

She believes that employee experience using AI at home and using personal projects can shape corporate policies as AI tools evolve.

She says there needs to be “a constant dialogue about which tools will help us.”