Dit is een paragraaf. Klik hier om je eigen tekst toe te voegen.
Dit is een paragraaf. Klik hier om je eigen tekst toe te voegen.
Dit is een paragraaf. Klik hier om je eigen tekst toe te voegen.
Dit is een paragraaf. Klik hier om je eigen tekst toe te voegen.
DIGIT
Pitfalls of Artificial Intelligence
Artificial Intelligence (AI) has rapidly emerged as a transformative technology, disrupting traditional industries and revolutionizing the way we live and work. However, as with any powerful technology, there are potential pitfalls and risks that must be considered. As AI continues to advance and become more ubiquitous, it is crucial to understand the potential risks and take steps to mitigate them.
In this series, we will explore 5 key pitfalls of AI. Every pitfall will be illustrated by a story and effective mitigation strategies will be discussed.
Pitfall 3: AI models are approximations and can be tricked
Illustrative story
To gauge the quality of education provided by universities, ranking models have been devised. These models output quality scores using inputs such as the number of papers published or the average salary of graduates of a university.
The number of papers published is a good indicator of the level of research but can easily be optimized by splitting one paper into five, publishing unfinished papers, or paying researchers to publish their papers under your university’s name. Likewise, the average salary of graduates can be optimized by only accepting rich students or by reducing the number of students in the philosophy and psychology faculties and increasing the number of students in the engineering and business faculties.
Universities today are more and more dependent on these rankings and will put great energy and effort to optimize the model. These optimizations do not always improve the educational quality of the university and put pressure on other universities to follow suit.
When applying an artificial intelligence model, the same can happen. We should be aware that once the model is in place, people will understand which drivers influence the model’s output and may seek to leverage this knowledge to their advantage.
Why?
Models approximate reality. IQ tests approximate intelligence, credit scores approximate individual financial solvency, and the GINI coefficient approximates inequality.
These approximations will never represent reality perfectly but are used to a great extent in society. When the inner workings of a model are public knowledge, it is easy to manipulate the model to predict a certain outcome.
All models are approximations. Essentially, all models are wrong, but some are useful. However, the approximate nature of the model must always be borne in mind.
George E.P. Box
How to mitigate?
AI models being approximations and being susceptible to being tricked can be a concern for companies. Here are some strategies that can be used to mitigate this risk:
-
Regularly update models: Changes in a model’s environment can quickly render a model outdated. Therefore, it is important to regularly update models with the latest available training data.
-
Implement security measures: Implement security measures to ensure AI models are robust and protected from attacks. Adversarial testing, a technique to simulate attacks from malicious actors, can be used to identify vulnerabilities in AI models.
-
Upscale in-house data literacy: Generate awareness around how to interpret and nuance AI models. Fostering a data-driven culture within your company will encourage collaboration and innovation around data analysis, ultimately leading to better company performance.
WANT TO KNOW MORE?
AI can bring tremendous value to an organization if it is well-managed and understood. However, implementing AI can be complex and time-consuming, requiring specialized knowledge and resources.
At BrightWolves, we specialize in providing customized advice and solutions tailored to specific business needs. Our expertise in AI can help accelerate your digital & data transformation by providing valuable guidance on best practices and implementation strategies.
What sets us apart is our focus on the business side of data analytics, rather than just the technical aspects. We understand that data is only valuable if it helps businesses make better decisions and achieve their goals.
If you want to know more, do not hesitate to reach out to our AI experts:
