AI employee monitoring has reached the gambling and live casino streaming industry. To protect game integrity, prevent fraud and keep track of service quality, employers are increasingly turning to artificial intelligence (AI) solutions. But when a casino operator deploys video analytics that track staff behaviour and physical condition, it can walk straight into a legal minefield.
In Latvia, too, employees have started raising concerns and filing complaints about AI employee monitoring in their workplace. So where does European Union (EU) law draw the line between a company’s legitimate interests and an employee’s fundamental rights?
In this article
Why casinos turn to AI employee monitoring
From management’s point of view, the wish to introduce smart monitoring tools is understandable:
- Game integrity. Fraud risks must be prevented, such as passing secret signals to players or other breaches of the game rules.
- Quality and safety control. Croupiers and game presenters need high concentration, so employers want to spot fatigue early, before it leads to mistakes.
- Regulatory requirements. In gambling, video surveillance of gaming halls and tables is standard and often a legal requirement.
These interests are legitimate, but the way they are pursued must stay within the limits set by law.
The employee’s perspective: privacy, dignity and stress
For a croupier, card dealer or stream host, working in front of a camera is nothing new. There is, however, a real difference between footage being reviewed by a person (a security officer) for safety purposes and an employee being continuously scanned and scored by an algorithm.
Once AI employee monitoring starts analysing facial expressions, micro-movements, blinking or stress levels, the employee faces:
- Intrusion into privacy. Physiological data and psychological state are deeply personal information.
- Psychosocial pressure. Knowing that a system measures every facial movement or sign of tiredness creates extra stress and a sense of algorithmic control.
AI employee monitoring under the EU AI Act
The EU Artificial Intelligence Act (Regulation (EU) 2024/1689) sets clear limits for employers.
Emotion recognition in the workplace is prohibited
Since 2 February 2025, it has been prohibited in the EU to use AI systems that infer the emotions of people in the workplace (Article 5(1)(f) of the AI Act). The only exception covers medical or safety reasons, and it is interpreted narrowly. Tools that try to read an employee’s stress, frustration or mood from facial expressions or voice are therefore unlawful in a casino workplace.
An important nuance: under the AI Act, emotion recognition does not cover physical states such as pain or fatigue. A system that only detects signs of fatigue is not automatically prohibited. The line is very thin, though: if the same system also assesses stress, attention or mood, it may fall within the ban. And fatigue detection is still subject to the GDPR and the high-risk rules.
High-risk AI systems
Even if AI is used only to analyse behavioural anomalies and dealing accuracy, AI employee monitoring that tracks and evaluates staff performance and behaviour qualifies as a high-risk AI system (Annex III, point 4 of the AI Act). It must meet strict requirements on risk management, data quality, human oversight and transparency, and the employer must inform staff and their representatives before deployment.
Following the Digital Omnibus amendments adopted in 2026, these requirements apply from 2 December 2027. Systems being rolled out now are worth designing to that standard from the start, so they do not have to be rebuilt later.
AI employee monitoring and the GDPR
The GDPR requires proportionality and data minimisation. The employer has to show that its objective (for example, game security) cannot be achieved by less intrusive means.
- Processing biometric data to identify a person is special category processing (Article 9 GDPR) and needs a separate legal basis.
- Employee consent is usually not valid in this setting because of the imbalance of power between employer and employee.
- Where the system itself takes decisions with significant effects on an employee, the Article 22 GDPR restrictions on automated decision-making apply.
- Systematic AI employee monitoring with new technologies requires a data protection impact assessment (DPIA, Article 35 GDPR).
On individuals’ rights after a data protection breach, see also our article CSDD data breach: do you have the right to claim compensation?
Latvian Labour Law and the Data State Inspectorate
Latvian Labour Law requires employers to ensure safe working conditions while respecting employees’ private lives. The Data State Inspectorate (DVI) consistently stresses in its practice that comprehensive and continuous monitoring of employees is not acceptable, and that employees must be clearly told how and why their data is processed. This applies fully to AI employee monitoring.
4 rules for lawful AI employee monitoring in casinos
Companies that develop or deploy game monitoring tools should remember one principle: the technology should focus on the game, not on the employee’s personality.
- Adjust the AI functionality. Drop features that analyse employees’ emotions, stress or facial expressions. Manage fatigue risk through sensible scheduling of working and rest time.
- Focus on objects, not faces. Point AI analytics at the gaming table, cards, chips and the logic of the game flow.
- Keep a human in the loop. Final decisions on breaches or suspicious activity should always be taken by a person, not automatically by an algorithm.
- Assess before you deploy. Carry out a DPIA and an AI risk audit before launch, and inform your employees.
AI employee monitoring: frequently asked questions
Is AI employee monitoring legal in an EU casino?
Yes, within limits. Systems that analyse the game (cards, chips, table activity) can be lawful if they meet GDPR and AI Act requirements. Systems that infer employees’ emotions are prohibited.
Can AI employee monitoring detect dealer fatigue?
Fatigue is a physical state, so detecting it is not automatically banned emotion recognition. However, the line is thin, and such a system still needs a legal basis, a DPIA and human oversight.
Do we need a DPIA before introducing AI employee monitoring?
Yes. Systematic monitoring of employees with new technology requires a data protection impact assessment under Article 35 GDPR before launch.
When do the high-risk AI Act rules apply to AI employee monitoring?
After the 2026 Digital Omnibus amendments, the high-risk requirements for employment-related AI systems apply from 2 December 2027. The ban on emotion recognition at work has applied since 2 February 2025.
Is your company planning or developing AI solutions?
The rules are changing fast, and the line between an innovative security tool and a serious breach can be very thin.
R|R Law and Finance advises on the EU AI Act, the GDPR and employment law, helping companies adopt new technology safely and lawfully.
Contact us for a legal review of your AI employee monitoring solution: www.rrpartners.lv