“The key problem that we’re facing now is that our dataset is evolving. And by that, I mean the schema is evolving all the time.”
We had a great chat with Adriano Soares Koshiyama, CEO of Holistic AI, about AI in risk management, Conversational AI and how technology is changing the way that we monitor, map and verify risk. Among Koshiyama’s achievements, he is also part of the Alan Turing Institute as an Enrichment Scheme Student, focusing mainly on Data Science, Machine Learning, Statistical Methods, Optimization, and Finance.
Trends in Conversational AI & Data-Driven Machine Learning
“We notice that there is a huge push for conversational AI. There’s a lot of new data-labeling companies coming out and there is this… kind of pendulum swinging more towards this data-centric approach to machine learning.”
“I think the pendulum will, in the next couple of years, move back to the model-centric… You have all these models that are becoming quite common… like BERT, GPT-3, etc. I’m sure that we are going to see way more of this popping up in different areas like computer vision, speech recognition, financial time series, etcetera.”
Market Growth & the Future of Artificial Intelligence
“It’s a growing industry… The long-term reason is because AI is being adopted more and more. As we can imagine, there will be a growing demand for auditing assurance, risk management… but that’s a long-term trend.”
“… The next two, three years, you’re going to see the EU making big interventions in the regulatory space. As similarly as they did with the GDPR, which was a big intervention in terms of data. They’re going to do the same intervention now with AI. So that’s going to become a big thing. And generally speaking, companies are very concerned already with risk.”
More Solutions for ‘The Big Data Problem’
“I still think that the data is the big problem. Even though I… tell people let’s stack it with all of our data – you have big data, you have data protection officers, you have data scientists, you have the whole thing… I think that kind of my sense is that maybe by the mid to end of this decade we will be really focusing more on the models.
“… It’s [like] capitalism in a way, right? I mean, when there’s a lot of value to be absorbed, or conquered, or captured – companies will move to that to try to solve that problem. And I think we’re in that moment that there’s a lot of startups and companies emerging.”
Prepare for more on big data, Ethics and AI in Risk Management
An episode you absolutely can’t miss!
Feel free to share any thoughts, comments or ideas with us!