Three quarters of Polish entrepreneurs already use artificial intelligence or plan to implement it. At the same time, 62% of business leaders believe AI makes it easier for criminals to develop new fraud techniques. The same technology that promises productivity gains is therefore becoming a new layer of operational risk.
Companies are using AI to automate routine tasks, analyse data, prepare content and identify suspicious activity. But generative systems also make it easier to produce convincing phishing messages, imitate communication styles and create fake documents or other material used in social-engineering attacks.
AI adoption is moving from experiments into daily work
The rapid spread of AI means that the discussion is no longer limited to technology departments. Sales, marketing, finance, customer service and administration are all finding practical uses for generative tools.
For smaller companies in particular, AI can reduce the cost of tasks that previously required specialist software or external support. That makes adoption easier, but it also means tools can enter an organisation faster than internal security rules are updated.
Fraud becomes cheaper to personalise
Traditional phishing often relied on poor language and generic messages. Generative AI changes the economics of fraud by allowing attackers to create polished, context-specific messages at scale. A fake request can be adapted to the language, tone and business context of a particular company or employee.
The risk becomes more serious when AI-generated text is combined with voice cloning, manipulated images or information collected from public sources. A message that appears to come from a manager, supplier or customer can look much more credible than older forms of spam.
Verification procedures matter more than detection by intuition
Companies cannot assume that employees will always recognise a fraudulent message from spelling mistakes or unusual phrasing. The more useful response is to build verification into processes: unusual payments, changes to bank details and requests for confidential information should require confirmation through a second channel.
This is particularly important for finance teams, executives and employees with access to customer or payment data. AI awareness training also needs to move beyond general cybersecurity advice and address realistic examples of impersonation and synthetic content.
The next stage is governance
For Polish companies, the next stage of AI adoption is likely to be less about whether to use the technology and more about how to use it safely. Organisations need clear rules on which data may be entered into external AI systems, who checks generated outputs and which decisions must remain subject to human approval.
The two survey figures capture the tension well. Businesses see enough value in AI that three quarters are already using it or planning to do so, while a clear majority simultaneously recognise that criminals are gaining access to the same capabilities.
The competitive advantage may therefore come not from adopting AI first, but from integrating it into processes without weakening security, confidentiality and accountability.
Source: survey data reported by CEO.com.pl.
