Technology By

AI-using employees now riskier than hackers for Czech firms

Worse than hackers: AI-using employees have become a threat to Czech companies

Artificial intelligence makes work easier, but it can also put a company in the firing line: according to Irena Hysková, head of the Czech cybersecurity firm Alintrust, employees often secretly use free public chatbots, such as ChatGPT, for routine tasks and feed them confidential company information — data that can later end up in the hands of hackers and make an attack on the company easier.

How chatbots help hackers

"If the AI tools a company uses are free, then all the information entered into them can be accessible to anyone else," Hysková explains. According to her, employees often use public chatbots in secret, without the knowledge of the IT department and without any security rules, simply to make their own work easier.

The expert notes that artificial intelligence can just as easily be exploited by a would-be attacker. "In the past, if someone was preparing a cyberattack, they needed to gather information about the company and its culture: who works there, what forms of address are used, and then draft a phishing email. Today AI helps do this very quickly, and even the employee themselves may not realize whether it's a cyberattack or not," Hysková adds.

A rising number of incidents

According to the National Cyber and Information Security Agency, the Czech Republic recorded 203 such incidents last year — about sixty fewer than in 2024 — though the agency warned that attacks are becoming more sophisticated. From the start of this year through the end of August, 159 incidents had already been recorded. Victims have included large companies as well, such as the Partners group. Hackers often attack not the company itself but its weaker link — its suppliers.

According to IBM, Microsoft and other companies, employees and unsanctioned AI will be the fastest-growing security threat of 2026. Surveys show that 80% of employees use unauthorized AI tools at work. Microsoft reports that 78% of employees bring their own AI tools to work, while KPMG, following a survey across 47 countries, warns that 57% of employees actively hide their use of AI from their employer.

What the expert advises companies to do

According to Hysková, management needs to clearly define which tools are allowed and will be used within the company, and employees need to be trained on what counts as sensitive data. "I strongly advise employees against entering anything at all into AI tools. Very often people thoughtlessly enter names, surnames, personal ID numbers, business contracts and a host of other things. These tools are helpful and boost productivity, but entering this kind of data into them is simply not worth the risk — the company could end up in the crosshairs of an attack," she says.

Hysková stresses that she does not think banning AI tools is the right approach — they bring real benefits and simplify routine tasks. But their use needs to be kept under control, and companies need to know exactly which tools their employees are using.

About Alintrust

Alintrust has operated under its current name since February 2026, when it split off from the former firm Thein Security — the financially healthy part of entrepreneur Tomáš Budník's troubled Thein group. Alintrust now operates as an independent cybersecurity company, with the J&T group remaining one of its major shareholders.

Alintrust's clients include major retailers, banks and pharmaceutical companies. Last year the firm's revenue topped CZK 386 million, while EBITDA grew 60% to CZK 30 million. The company plans to grow further through acquisitions — according to Hysková, Alintrust has its eye on two or three companies that could round out its cybersecurity services portfolio.

Around Prague, based on reporting by seznamzpravy.cz

Share: Telegram WhatsApp

More science news

How a $5 Billion AI Push Is Reshaping Which U.S. Science Gets Funded
While funding cuts and delays have hit public health and climate research during President Donald Trump’s second term, artificial intelligence and quantum science have found firmer footing, Science Ne
Young Mathematicians Are Using AI Daily, and Most Say It Is Making Them Rethink Their Careers
Nearly every early-career mathematician who responded to a new survey says they use artificial intelligence, yet many feel uneasy about what it means for their future, Scientific American reported. Se
How Did a Field of Pure Sulfur Rocks End Up on Mars? Two New Studies Offer Answers
Scientists now have two possible explanations for a puzzling discovery on Mars: a field of rocks made of nearly pure elemental sulfur. NASA's Curiosity rover first found them in 2024, according to Sci
BMI Alone Missed Obesity in Up to 8.6 Percent of People in a Study of 844,108 Adults
Body-mass index may misclassify obesity for a meaningful share of people, according to a new international study published in Nature. ScienceAlert reported that the researchers compared BMI with a bro
Trump’s Asia Trip Ends With First Summit Among Him, Xi and Putin
President Donald Trump heads to Asia this week for a trip that will end with an unprecedented three-way summit with Chinese President Xi Jinping and Russian President Vladimir Putin, according to The
Sponge-on-a-String Test for Esophageal Cancer Risk Recommended for NHS Use
Thousands of people in England who are at higher risk of esophageal cancer could soon be offered a seven-minute test on the National Health Service that involves swallowing a small capsule on a string

Related news

Microsoft has restored access to its flagship software more than 24 hours after a global outage cut off Outlook, Teams and the Copilot assistant for customers, according to Stockhead. The company said a faulty update was to blame, not a cyb
Coles, one of Australia’s two big supermarket chains, said Monday it is investigating a cybersecurity incident. For shoppers the practical message so far is that stores and online ordering are running as normal, and the company says it has
Alibaba has unveiled Qwen4-Max, its new flagship artificial intelligence model, saying it handles complex professional and scientific tasks better than its predecessor while adding stronger protections against safety risks, according to The
Former Nvidia chief Jensen Huang, who resigned on Thursday, is considering a return. According to the Financial Times, a group of executives is forming in Silicon Valley to lobby for his comeback and to seek the White House's support.Allies
A rare close approach of Jupiter and Mars will take place over Czechia overnight from Sunday, Nov. 15, to Monday, Nov. 16: the planets rise above the horizon around 11 p.m. and stay visible until dawn. The closest approach comes at 3:06 a.m
The Chinese electric aircraft COMAC CE-25A took to the air with people on board for the first time on 14 November. The test took place in Binzhou, in Shandong province, with two pilots and a flight-test engineer in the cockpit. The crew is
On Saturday, SpaceX caught its Starship spacecraft for the first time after it returned from orbit: at the Starbase site in Texas, a special mechanism on the launch tower grabbed it. The Super Heavy booster could not be saved, so the missio
The global Microsoft outage, which began on Thursday, 12 November and lasted more than 24 hours, has left customers with an uncomfortable question about money. Compensation is available, but only as a credit toward the cost of the service i
Italy's data privacy authority, known as the Garante, has fined Meta 805 million euros ($887 million) for accessing photos and videos on Android users' phones without their specific consent. The regulator announced the penalty Saturday, acc
A campaign dubbed wfx-loot has used thousands of hijacked GitHub accounts to slip credential-stealing code into software projects, and researchers have confirmed that the malware ran and stole secrets from at least 1,835 repositories, accor
Microsoft has suspended the rollout of all new updates, is moving part of its traffic off the affected infrastructure and is restarting servers to restore Microsoft 365, Outlook and Teams. “We are working intensively on every possible way t
Rockstar Games’ Grand Theft Auto VI is set to be released worldwide on Thursday, Nov. 19, for PlayStation 5 and Xbox Series X/S, with a PC version to follow at a later date, according to the Tullahoma News. The standard edition costs $79.99