Artificial Intelligence

Trainings

A comprehensive set of practical trainings to master AI & Data quality

Learn more

READINESS ANALYSIS

Understand your organisation’s AI quality readiness level

Learn more

GUIDED ASSESSMENT

A systematic approach to assure AI quality and compliance

Learn more

 

WHAT IS AI QUALITY

The quality of an AI system is its property to meet clearly defined criteria. The quality criteria of an efficient ecosystem must be:

• Transparent
• Measurable
• Auditable

AI quality refers to the quality, during development and operations, of:

• AI system, such as (but not only) accuracy, reliability, and performance, including infrastructure
• AI Data, its lifecycle and how it relates to the AI system.
• AI governance, covering for example compliance, strategy and human resource
• AI Process management, all along life cycle and risk management, such as cybersecurity

 

WHAT ARE THE PAIN POINTS WE COVER

The quality of an AI system is its property to meet clearly defined criteria. The quality criteria of an efficient ecosystem must be:

• Regulation / compliance (Risk)
• Responsible use of AI (ethical AI)
• Scalability (internal / Governance)

 

WHAT IS THE TÜV SÜD COMPREHENSIVE AI QUALITY FRAMEWORK 

It covers the whole life cycle of AI and data, split in 6 pillars and analysed with standards, regulation and best practice.

Why Choose TÜV SÜD As Your AI quality Partner

  

AI Expertise

Combining our AI competence and expertise with AI, we have developed a uniqueAI Quality Framework. It consolidates all relevant existing and upcoming standards and regulations to expedite your adoption of AI & Data quality.

 

Committed to Quality 

Inspiring trust in technology has been our DNA and has served us as a guiding principle since our inception. Translating this into today’s digital world, we ensure the safety, security, reliability, and robustness of AI systems.

 

 

Independent and Impartial

As an independent and impartial organization, we have been delivering testing, inspection, and certification services, for more than 150 years.

 

We require your consent to load the YouTube video. This service uses cookies that are set when loading and personal data may be transmitted to third-party platforms and unsafe third countries. To enable the service and to give consent, click on ‘Manage cookie settings’ and enable targeting cookies, then proceed. You can withdraw your consent at any time using the cookie hover. The withdrawal of consent does not affect the lawfulness of processing before a withdrawal. For details, please refer to the data protection information.

We require your consent to load the YouTube video. If you consent, external content will be displayed, and your personal data may then be transmitted to third-party platforms and unsafe third countries. To enable the service and give consent, click on “view video”. You can withdraw your consent at any time using the cookie hover. The withdrawal of consent does not affect the lawfulness of processing before a withdrawal. For details, please refer to the data protection information.

• Is EU AI Act relevant for me?

  There is no straight answer to this, but if you are using or programming AI being in touch with Human, then most probably yes. There are many other situations where you may be concerned: the EU AI Act is a risk-based approach. Contact us to clarify your situation. 

• How difficult is it to implement a AI quality management process?

  This is a journey across your whole organisation, your AI system, your data and your processes. This is about learning and implementing how to make AI right: processes, documentation, governance and infrastructure. Depending on where you stand on your digital transformation, this can be demanding. However, no matter what your needs, we are here to help you understand your situation.

• We have AI in production, and we get good results. Why do we need extra AI quality?

  Let’s use an analogy. You are a car manufacturer; you are happy with your car specifications and proud of how fast you accelerate from 0 to 100km/h. Does this mean you would produce this car without compelling quality procedures? Do you know how to maintain it? What happens by freezing temperatures? Your CTO is leaving, and you are lost? Your insurance is asking you for proof of compliance? You are facing adversarial attacks,...

  You understand now what this is all about: AI touches your entire organisation, and needs to be addressed as such.

• Do I need AI Quality for IOT devices?

  Sure! AI is not only to be managed in your production or data science department, but as well when embedded into IOT devices. You will want to ensure that your customers and final users are dealing with a trustworthy AI, wouldn't you?

• Can AI Quality protect me from liabilities?

  This is one of the major reasons to implement AI quality framework, based on standards and regulations: you will tremendously reduce the risk of liabilities, as they just are very unlikely to occur. And shall they nonetheless occur, your compliance to legal requirements and standards will be a really strong umbrella facing liabilities issues.