by , on Oct 08, 2019 11:54:39 AM

Malta is taking up multiple initiatives, as part of the Malta AI Taskforce Strategy, to provide the legal framework to help implement ethical and trustworthy AI in Malta. In support of this Strategy, the MDIA has issued four consultation documents, as highlighted below, which set up the certification process for AI systems.  

AI Innovative Technology Arrangements Consultation Documents

The Malta Digital Innovation Authority (hereinafter referred to as ‘MDIA’) has issued four consultation documents which help pave the way for the development of the AI Innovation Technology Arrangement certification framework (hereinafter referred to as ‘AI ITA’). The most salient features that an AI ITA should include are highlighted in the documents, along with the process of certification and the components which make up an AI ITA.

These documents are as follows:

• AI ITA Blueprint Guidelines
• AI ITA Nomenclature
• AI ITA Guidelines
• AI System Auditor Control Objectives

 AI ITA Blueprint Guidelines

The AI ITA Blueprint Guidelines mainly concerns itself with the application stage of the ITA certification. It sets out the requirements needed for the application to be accepted by the Authority. Such requirements that need to be satisfied include the following:

• A high-level description of the scope and purpose of the AI ITA;
• A detailed description of the technical specifications of the AI ITA, along with details of the AI engine and data processing analysis, among other things;
• Details on the forensic node as may be required in the Forensic Node Guidelines;
• Information security and risk assessment requirements;
• Powers and technical features available to the Technical Administrator;
• Terms of service and other useful information.

AI ITA Nomenclature

The AI ITA Nomenclature focuses on providing an overview of the processes and what makes up an AI ITA. This determination should be done through the descriptions as required by the Blueprint Guidelines. The Nomenclature Guidelines focus on the following processes:

• Data collection and storage processes;
• Data processing and analysis processes;
• AI engine processes;
• Application and implementation;
• ITA harness;
• Variations.

 AI ITA Guidelines

The AI ITA Guidelines apply to any ITAs seeking voluntary certification by the MDIA, and sets out some high-level standards which need to be followed by ITAs seeking such certification. These high-level standards include:

• Qualifying features and qualities;
• Eligibility standards;
• The application process:
• Stage 1;
• Stage 2;
• ITA processing fees;
• Other applicable documentation;
• General and specific requirements:
• Fit and properness;
• Positive assurance from System’s Auditors;
• Appointment of Technical Administrator;
• Adequate disclosures to users and affected third parties;
• Resident agent;
• Outsourcing;
• Certification;
• Ongoing obligations;
• Renewal process.

AI ITA Systems Audit Control Objectives

The AI ITA Systems Audit Control Objectives highlights the role of the Systems Audit as an assurance engagement, according to the MDIA’s AI ITA Systems Auditor Guidelines. Among the Systems Auditor’s roles, the Systems Auditor may clarify or amend appropriately any of the objectives listed in these Guidelines upon consultation with the Authority.

Conclusion

Through the creation of a certification process, and through the setting up of standards which AI ITA systems need to satisfy in order to be certified, the MDIA is thus ensuring that the AI ITA systems meet the requirements on an ongoing basis. This shall therefore provide for the further development of the AI industry in an ethically-conscious and trustworthy manner.

 

If you want to know more, contact us on [email protected] or visit our website on www.novo.legal to keep up to date with the latest industry news, both local & international.