The role of AI and Machine Learning in Risk Management

In this week’s blog post, we’re sharing insights on our interview with Terisa Roberts. She is a Director and Global Solution lead for Risk Modeling and Decisioning at SAS. She has extensive experience in quantitative risk management, and advanced analytics. She has worked in a variety of industries, including financial services, telecommunications, energy, retail and government. She advises banks and regulators around the world on best practices, topics in risk Modeling, Decisioning, and the responsible use of artificial intelligence and machine learning. She also holds a PhD in operations research and informatics.

COVID-19’s Impact on Risk Models

SAS has been an analytics leader for a long time. It has been named one of the top risk and compliance technology providers by Chartis for 15 consecutive years. It ranked fifth overall in the Global Chartis RiskTech100® 2021 report and won three industry solution categories with Model Risk Management being one of the three. As a leader in providing world-class analytical solutions for a wide range of business applications, they also include risk management, fraud detection and customer diligence services as well. They have recently started to move their solutions to SAS Wire which is a cloud native platform, opening a new range of opportunities for innovation.

But even the world leader companies were not safe from the disruptions COVID-19 has brought in. Risk Models have also taken a big impact with COVID. The Pre-COVID saying “Essentially, all models are wrong, some are useful” has now transformed into “Essentially, all models are wrong, some are useless” The reason for that is because the model inputs have moved outside of the calibration range of these models. The models have never seen the narrative of a pandemic play out, such as the global scale lockdown measures, unemployment, movement restrictions and so on . While we as individuals have been massively disrupted, so have the business models of companies alongside us.

From a risk management perspective, of course not every industry was affected on the same scale, but nevertheless they were all hit by the unpredictability and uncertainty towards the recovery path. Both the traditional statistical and econometric models have shown a breakdown, as well as the newer algorithms such as machine learning. As both models have had trouble coping, financial institutions had to apply their own human judgment and make adjustments and management overlays with those in place for the models.

AI and Machine Learning in Risk Management — Benefits and Disadvantages

What we see with the use of AI and machine learning for risk management, it’s really good at jobs that require repetition and that can be automated. So AI and machine learning is good at going through volumes and volumes of unstructured and structured data and picking up patterns that the human eye might miss. So possibly it’s less glamorous than what we see in the movies, but definitely it’s starting to pay off and deliver tangible benefits in the areas of credit scoring. For example, AI would be a big asset to be able to offer credit and access to credit, to minority groups where perhaps the credit Bureau funds are not available or in the areas of fraud detection and cybersecurity.

Ofcourse, for many businesses, AI and machine learning are still complicated processes and not fully implemented. In the area of risk management, Terisa believes that there’s still a lot of caution in the industry, as technologies with more complex and sophisticated algorithms are commonly associated with known risks such as lack of transparency. It’s not so easy to explain to the various stakeholders how the model works but there’s also great strides being made in addressing the explainability of all of these algorithms.

Another challenge that is often associated with AI and machine learning is that it may perpetuate and amplify a bias that might be present in training data. There might be some individual and societal biases in the data that these algorithms are not smart enough to adjust for and amplify the bias. We’ve seen some examples of this in the news, where facial recognition systems and in the credit space and approvals of credit cards, some big technology companies gave women a much lower credit limit because of the historical data. Although a bias is an important factor to take into account, definitely it is a measure to be kept in mind as it can lead to wrong calculations such as these. Once these are properly addressed however, AI and the machine learning models might even help us to make more consistent data-driven decisions.

Existing Models and Integration of AI and Machine Learning

There’s always a question whether the existing model governance frameworks are sufficient for AI and machine learning or whether there should be new ones developed and approved by different bodies or if AI should validate another AI. Probably the best course of action would be to add layers into levels of validation, to make room for AI and machine learning and focus on validating feature engineering as well as other hyper parameter tuning methods. These are definitely putting a lot more demand on validation teams and model governance, meaning that integration of AI and machine learning doesn’t mean such functions are going to be obsolete.

Taking a Strategic Approach in Technology and Risk Management

The process that takes to put models, whether they are traditional or more innovative into production is quite long, averaging from six to nine months for the financial services industry. This means that, for banks to take new innovation and modeling seriously, they would need to rethink the model life cycle process, and look for efficiency gains perhaps in standardizing the data that goes into the models as well as looking at the deployment. Ideally, a traditional model might have a handful of input variables in comparison to an AI or machine learning model that can easily turn into hundreds and thousands of input variables. So if you have a manual process of deploying those models into your decision architectures, that might not be scalable for you as the models increasingly grow in numbers.

Understandably, there’s a big hype around robots and computer vision and how they are able to replace humans in the future. Nevertheless, there is more to it than how we perceive it in science fiction and that’s not going to be the case for a long while. To be able to take a strategic approach, we need to break the hesitation and embrace new technologies. Looking at not replacing all your traditional models with machine learning but looking at where the AI and, and the machine learning model can give you additional value is a big asset. If we look at the science fact, there are areas where innovation can make a real difference, as we see it in customer experience. For example, by having these models to estimate income, you can gain efficiency and get much more accurate data.

Takeaway Points

First major takeaway point is to rethink, innovate and implement. When it comes to AI and machine learning, we need to rethink our architecture, the infrastructure behind the data management and the modeling. We should be looking into the deployment of the models and making sure that the infrastructure is future-proof to handle the sophistication of these models and look for areas where the manual processes can potentially be improved with automation, for those efficiency gains and scaling because of the operational efficiencies.

Second major takeaway point is to collaborate and contribute. There are a lot more emerging risks, such as the climate risk or geopolitical risks. So as a community, as the front representatives of the industry, regulators, risk management professionals, technology providers should be collaborating on the solutions that are practical and help everyone to operationalize more of the technologies. This will create a better world, a safer place, and also more equitable and fair outcomes from our risk decisions.

Closing Words

For now, this sums up the key points of our interview. As the Global Risk Community team, we once again thank Terisa Roberts for her insight on AI and Machine learning in Risk Management. More information about this topic is available in our original interview, which is accessible here.

#risk #AI #Machinelearning #management #finance