John Ascroft Tue, Jul 25, '23 11 min read

Exploring the Frontiers of AI: Insights from an AI Roundtable at Jade Software

Jade Software recently held a roundtable session with customers and staff to discuss the latest developments in the world of Artificial Intelligence (AI). We were fortunate to be joined by Professor Richard Green, who heads the Computer Vision Research Lab at the University of Canterbury. Professor Green has extensive experience and interest in the application of AI, particularly in areas involving computer vision and robots. Greg Smith, Jade's Head of Architecture, also provided valuable input for the session. 

AI has been making headlines lately, largely due to the publicity surrounding ChatGPT. This technology has not only made AI interesting but also accessible to a wide range of people for general-purpose usage. It can be likened to the Netscape of AI, offering a user-friendly view into a highly technical world that previously required specialised knowledge to understand and utilise. The widespread availability of this window into AI capabilities has naturally generated a lot of hype.

It is important to realise that AI has actually been around for a long time and has been progressing in terms of its capabilities. There are numerous existing applications of AI, such as healthcare, where it assists in analysing radiographic images, or finance, where it detects fraud and anomalies in transactions, among many other applications.

Professor Green and his team at the University of Canterbury have been at the forefront of combining AI with Robotics, particularly in the field of Agriculture. Some of their recent successes include autonomous drones for weed spraying over farmland and underwater operations supporting aquaculture. These applications merge robotic capabilities with computer vision to understand what needs to be done and execute the tasks. For example, the weed spraying drone can be programmed to fly over a farm area, identify weeds and wilding pines, and precisely spray them while leaving desirable crops and plants untouched.

During our discussions, Professor Green and Greg Smith explained the distinction between what they refer to as 'Small AI' and 'Big AI'.

'Small AI' can be characterised as applications that involve getting computers (and robots) to perform intelligent tasks (like the University robots) or creative generation. Examples of creative generation include text generation, as seen with ChatGPT, or code generation, as with Co-Pilot.

On the other hand, 'Big AI' refers to building human-level intelligence, also known as Artificial General Intelligence. This is an ambitious goal, and it is widely agreed that we are still many years away from achieving this capability. There is also a consensus that this form of AI will require firm safeguards to protect mankind, and we should be researching and building those safeguards now, long before they are actually needed.

Professor Green pointed out that although we already have driverless technology enabling vehicles to follow routes and drive on well-constructed roads, building cars without steering wheels that can go anywhere poses a significant challenge. Even achieving this goal is not currently within sight due to the added complexity of dealing with edge cases that humans can handle based on experience and assessment.

During the roundtable, we also discussed the application of AI to computer fraud and scams, both from the perspective of detection and as an aid in perpetrating them.

Most fraud and money-laundering detection applications now incorporate some form of AI and machine learning to detect and report anomalies and potential fraudulent transactions. This is an ideal application, given its well-defined target space, abundance of data, and the needle-in-a-haystack nature of the problem.

Professor Green also highlighted that international scammers now utilise AI to generate plausible and targeted phishing emails that have a higher chance of tricking the recipient. These criminals and hackers behind such attacks are increasingly sophisticated. In cases of ransomware and penetration, the actors involved are sometimes government-sponsored and leverage every possible tool.

Most of us attending the session could envision AI being applicable to our respective fields, and it will be intriguing to witness how the technology develops and the changes it brings in the coming years. One thing is certain: AI will not render everyone jobless in the foreseeable future!

Talk to us about making data meaningful