Transparency and Trust in Artificial Intelligence
Artificial intelligence is revealing the hidden value of data by automating decision-making. But poorly designed AI solutions can spiral out of control. There is, therefore, a growing need for artificial intelligence that can be understood and explained, operates according to regulations, norms and values, and is comprehensible and acceptable for customers and the public.
AI is automating decision-making:
AI solutions spiraling out of control:
Opaque fraud detection
The Dutch government’s risk indication system known as SyRI wrongly identified people who (potentially) were at higher risk of committing social benefit fraud, with devastating consequences for the individuals affected. SyRI was later deemed illegal by the Dutch court because of a lack of transparency, while public outcry about the secretive AI decision process forced the responsible State Secretary to resign.
Sexist credit cards
Tech giant Apple became the centre of a (social) media storm when it was revealed its credit risk AI system deemed men far more creditworthy than women. Even Steve Wozniak, one of Apple’s co-founders, tweeted that he had been given ten times the credit limit of his wife despite the pair having no separate cards, accounts, or assets.
AI predicts but does not prevent maintenance
Prevention is better than cure, and not knowing why an AI predicts maintenance does not enable you to prevent it. Being unable to understand why an AI schedules maintenance at best leads to symptomatic treatment, while not tending to the root cause, costing time and money over and over again.
Racist robotic judges
The COMPAS algorithm that informed judges about the risk a defendant would re-offend, helping them to decide whether to release or detain a defendant before trial, turned out to discriminate against black people. The algorithm was nearly twice as likely to incorrectly predict that black defendants were at “high-risk” to re-offend compared to white defendants, whilst white defendants were more likely than black defendants to be incorrectly flagged as “low-risk”.
The need for a trusted solution
Organisations, governments and the public have concerns and demand transparency and AI they can trust.
To take full advantage of AI you need a solution that can be understood, explained, and that operates wisely. It’s all about building trust.
FACT-AI: Our framework for responsible Artificial Intelligence
At VIQTOR DAVIS we identify four key drivers to build trust in AI.
FACT-AI approach for responsible AI
At VIQTOR DAVIS we have developed a state-of-the-art approach to implement our FACT-AI framework throughout an AI’s lifecycle.
We design, develop, deploy, and sustain your next AI solution using our FACT-AI approach utilising state-of-the-art methods and techniques to build trust.
We assess existing data and AI driven processes, products or services, facilitate interpretability of models, and detect and mitigate unfairness to build trust.
Explaining the blackbox in automated credit risk assessment
In the financial industry there is a growing demand for simplifying and speeding up processes such as replying to credit applications. At the same time, regulators require transparency about decision-making. VIQTOR DAVIS helped an international Global Vendor Company to these demands by combining the predictive power of blackbox AI with our AI explainer module. By opening the blackbox, our solution provides highly accurate predictions and transparency of decisions to all stakeholders.
- 90% of applications handled automatically without an increase in risk exposure.
- AI recommendations are interpretable for business users.
- AI decisions are comprehensible & acceptable for all stakeholders.
Fair recruitment with our AI-matchmaker
The recruitment market is under pressure to automate its processes because of the demand for qualified candidates, growing recruitment costs and increasing regulation. VIQTOR DAVIS can help organisations measure and mitigate unfairness within AI recruitment solutions.
VIQTOR DAVIS offers a unique combination: We are your partner in data strategy, governance, management, science and analytics. We provide professional services, full-service solutions and knowledge transfer. We call this Data Craftsmanship. Build trust in your AI solution? Get in touch.
FACT AI White Paper
We have written a White Paper about this subject. Download the document instantly.
FACT AI White Paper
pdf | 1.372 MB