The future of generative AI in detecting financial crime
Much attention is being given to Chat GPT—it’s seen as readily available, generative AI that can assist us in daily functions. Need an essay for school or an article for an upcoming deadline? A quickly written resume or cover letter? A new song for your oldies cover band? Or maybe you’re just in need of a joke or a chat companion.
But the promise of generative AI goes beyond personal needs, and many are already exploring the business use cases. Trial judges in Columbia and India recently made news by handing down rulings based on advice given to them by Chat GPT. Samsung employees used it to debug code (and got into trouble for it). Fashion designers are using generative AI to design sneakers. Adobe added it to Firefly for graphic designers to generate images. And many more uses are coming. It’s a race to use it in helpful, secure ways.
You might be thinking, “What’s in store for generative AI in financial crime detection and compliance?” First, we need to consider what it’s good for and what its limitations are.
Generative AI admittedly does have some limitations, just like all other AI models. It requires huge amounts of training data to accurately learn patterns and detect anomalies. On its own, it may have trouble detecting changing patterns. Because it learns from data on which it is trained, its responses could contain biases.
But well-designed tools that use generative AI can provide more value than these limitations might suggest. Generative AI can create new summaries and insights on top of existing datasets. This function is potentially very powerful for financial crime investigators and regulatory compliance teams.
How generative AI will shape the future of financial crime detection
It’s fair to say that generative AI, like other forms of AI, will find its way to risk and compliance systems. When it does, financial crime teams that integrate generative AI into their workflows will benefit greatly.
Generative AI will be used to arm financial investigators with more powerful tools to fight bad actors. We know that the volume and complexity of data that investigative teams face is staggering. Generative AI could decrease that cognitive load by acting as the liaison between the investigator, systems and data stores they need, streamlining the investigative process.
A system that uses this technology could simplify data searches, information mapping and narrative composition, allowing the investigator to focus on what counts—making decisions. Generative AI will also provide a useful and natural interface for what-ifs, so investigators can dig deeper when needed. Generative AI could also be used as a training aid for risk teams, and to ensure consistent summaries and narratives.
Generative AI has a bright future in our industry…a future isn’t that far off.