The Future of Fraud: The latest on AI
This new technology brings a whole “can of worms” along with it for the Credit Management and Debt Collections Industry. In this blog, we will highlight some of the advancements made and the increase of risk that that it brings with it and what it may mean for the future of fraud.
Progress in Artificial Intelligence
But first, what is GPT-4o?
If you have somehow not seen the viral videos making the rounds on social media, Open AI’s showcase blew everybody’s minds with the advancements it has made in the AI field.
If you have an Amazon Alexa that you use at home, imagine that but improve it by ten, a hundred or maybe a thousand times.
The showcase displayed GPT-4o’s advancements in Voice Variation, which sounds increasingly life-like and human, with intonations in the voice and fluent use of language as well as being capable of Visual Recognition, which demonstrated GPT-4o’s almost-human reaction to meeting Bowser, a small dog.
We have linked some of the showcases below, which display the Visual Recognition and Voice Variation technology in action and highly recommend that you watch them.
- Say hello to GPT-4o (1:23)
- Dog meets GPT-4o (0:27)
- Be My Eyes Accessibility with GPT-4o (1:07)
The potential effect on Fraud
The first thing that comes to mind following GPT-4o’s showcase is the implications that it has with fraud. One example OpenAI demonstrated was how interactive it can be in a virtual meeting setting. We’ve seen how far AI has advanced in less than 12 months Where will it be in the next five to ten years? As progress is moving quickly towards increasingly lifelike and realistic interactions with AI, how confident can we be in distinguishing it from a human and what implications will this have in fraud prevention?
Additionally, if you watch Figure’s latest video, the AI Robotics Company utilising OpenAI’s technology, you can see where the technology was at a few months before the GPT-4o, which unveiled a robot completing commanded tasks and reasoning with an “Uncanny Valley” human-like voice.
Without a doubt, GPT-4o still has a hint of that “Too good to be true” and “Too Perfect” nature about it for us to be able to not recognise that it is AI. However, that is not to say that voice variation, visual recognition and responses won’t be vastly improved to the point that the lines blur even more.
As Jamie Radford said in his podcast with the Chartered Institute of Credit Management, we’re increasingly going to need to have the knowledge, skills and experience to put our hand up and say that “something isn’t right here” when we’re dealing with potentially fraudulent activity.
Have you got the skills in place to be able to identify fraudulent activity? And do you think you could identify something at this level?
In our latest podcast “CEO to CEO: Two sides of the Same Coin”, Sue Chapple FCICM, CEO of the Chartered Institute of Credit Management mentioned that fraud was just “getting cleverer and cleverer”, which, following this showcase, we can’t help but feel that it certainly will.