Goal

High Quality AI aims to address the quality challenges that come with the current AI race. AI systems should be safe, robust, understandable, verifiable and auditable. AI systems should also be privacy proof, bias-free and fair.

7,5 bn

is the expected amount spent on intelligent process automation such as automated decision making in 2018

70,000

unemployed lost their welfare payments due to a rogue algorithm in Sweden

CHALLENGES

OPPORTUNITIES

Safety

Is the AI system robust? Is the algorithm properly programmed? Is it crash and hack-proof? Does the system do what it was intended to do? Is it effective? How do we ensure that it behaves safely in normal and unknown, critical or unpredictable situations?

Competitive advantage

AI that is provably safe, robust, cybersecure, bias-free, fair, explainable and verifiable will give companies a competitive advantage over their competitors that do not follow critical and rigorous quality standards. High quality AI will be a selling point.

The “black-box”

Currently, many AI systems are very difficult for users to understand. This is also increasingly true for those who develop the systems. In particular, neural networks are often “black boxes”, in which the (decision-making) processes taking place can no longer be understood and for which there are no explanatory mechanisms.

Risk reduction

Building and deploying high quality AI reduces the risk of liability claims and expensive recall actions as well as protects a company’s reputation. AI systems fall within the scope of the Product Liability Directive (according to the European Commission).

Privacy

Products with built-in AI such as household appliances, children’s toys, cars, health trackers transmit (often personal) data to the cloud-based platforms of their manufacturers. Trade in data is booming, and data is being used to profile people for various reasons. The upholding of the GDPR will be a challenge in the AI-era.

Growth and job creation

High Quality AI will lead to growth and thus job creation. New activities will emerge and new skills will be required to make sure the quality of the AI systems is guaranteed and continuously monitored throughout their lifetimes.

Bias

The development of AI is predominantly done by young, white or asian men, with the result that (whether intentionally or unintentionally) cultural and gender disparities are being embedded in AI. Also, the training data should be accurate and of good quality, diverse, sufficiently thorough and unbiased.

Responsiblity

Building, deploying and using High Quality AI will show an organisation’s  sense of responsiblity for avoiding adverse effects of low-quality AI-systems, that were trained with low-quality or inadequate datasets.

“High-Stakes algorithms
should not be crappy”

Cathy O’Neill

Theme In Action

  • RESEARCH

    Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nunc gravida lorem sapien, ut tincidunt dolor scelerisque eu. Etiam mollis urna euismod sollicitudin gravida.

  • ROUND TABLES

    Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nunc gravida lorem sapien, ut tincidunt dolor scelerisque eu. Etiam mollis urna euismod sollicitudin gravida.

  • CO-CREATION

    ALLAI will set up multi-stakeholder developper teams to ensure co-creation of AI that benefits both developers, users and workers alike, and respects social, ethical and legal implications.

  • MEETUPS

    Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nunc gravida lorem sapien, ut tincidunt dolor scelerisque eu. Etiam mollis urna euismod sollicitudin gravida.

CONTACT

ALLAI
Herengracht 247
1016 BH Amsterdam
welkom@allai.nl