EU AI Act Compliance Statement
The European Union Artificial Intelligence (EU AI) Act aims to regulate all AI systems marketed within the EU. It sets out four tiers of risk associated with potential AI implementations, ranging from Unacceptable to Minimal.
The SAI Platform is currently classified as a Minimal Risk system, and therefore there are no special requirements or obligations placed on the company, nor any of its customers.
ABOUT SAI GROUP
As a leading provider of computer vision technology to the retail industry, SAI Group is committed to developing AI within a comprehensive ethical and security framework. Shops are public spaces and therefore our technology must provide value to the owners while respecting the privacy and security expectations of their customers.
Our solutions analyse camera feeds to:
- improve store safety by identifying trip and slip hazards and alerting to aggressive or violent behaviour
- increase store efficiency by identifying gaps and highlighting service opportunities
- grow profitability by removing opportunities for shoplifting, and preventing misuse of checkout systems
The platform’s outputs are sent to store colleagues to focus attention on immediate actions, such as cleaning a spill, re-stocking a shelf, or collaborating with colleagues to prevent a theft.
The AI landscape remains fluid both in terms of technical capabilities and emerging legislative frameworks. As a company, we will continue to monitor the situation and ensure that its products and practices remain compliant in all territories around the world.
The EU AI Act
Artificial Intelligence is a very broad term covering dozens of disciplines and techniques. Rather than attempt to legislate for particular technologies or use cases, the European Parliament chose to base its approach on the possible risk of harm that may emerge.
Risk Classifications
The EU AI Act defines four risk categories.
Unacceptable Risk
AI systems violating fundamental rights or causing the manipulation or exploitation of humans are banned within the European Union.
High Risk
Systems that aim to exercise control of human rights or public spaces, such as vehicles, building access, legal decision making, and migration and border control are considered High Risk and are subject to strict mandatory requirements.
Limited Risk
Systems such as chatbots or emotion recognition systems are defined as having Limited Risk. These are permitted as long as the user is made aware that they are interacting with an AI system.
Minimal Risk
The majority of systems currently covered by the legislation are classified as minimal risk and are not made subject of new requirements or restrictions.
General Purpose AI (GPAI)
The Act defines GPAI as a system that can serve a variety of purposes directly, or when integrated with other systems. Very large models (1025 FLOPs and higher) must seek approval from the European Commission.
Sanctions
Maximum sanctions for infringing the terms of the Act are fines of up to 7% of a company’s global annual turnover.
SAI AND THE EU AI ACT
SAI Group Limited is a Provider of AI systems , according to the terms of the Act . We do not allow downstream modifications to our systems .
Our systems are marketed and deployed within the EU but do not meet the generality requirements to be considered GPAI, nor do they fall under any of the ‘Excluded ’, ‘Prohibited ’, ‘High Risk ’ or ‘Transparent ’ system categories defined in the Act.
In respect of Article 95 of the Act, we will publish a voluntary Code of Conduct on our use and implementation of AI in due course.
About SAI
As a leader in computer vision technology, SAI Group delivers cutting-edge, multi-modal AI solutions into retail environments. Using a unique platform approach, its technology uses existing camera systems to target losses, increase store safety, and underpin operational efficiencies.
All solutions are built from the ground up to ensure the highest levels of security and data protection, respecting the privacy expectations of the public and operating to stringent ethical standards while delivering substantial value to our clients. Globally, SAI monitors millions of transactions per day, protecting the revenues from tens of millions of product sales and hundreds of millions of customer interactions. Its models also accurately identify anti-social behaviour, aggression and violence, helping to de-escalate situations with real-time interfaces to security officers and operations centres.