Small and medium businesses in Australia are using AI more and more every day. A recent study of 400 Australian businesses by QBE into AI adoption found that 85% were already using AI in their operations.
While most surveyed businesses expect AI to have a positive impact, half had experienced a cyber incident in the last year and about two-thirds were concerned with AI-related cyber threats.
In this article we explore some of the ways industry has taken up AI, the risks associated with it and tips to protect your business.
What is AI
AI is used across industries in everything from the chatbot on your clothing website to identifying prospective geological areas for mining minerals.
It’s an umbrella term for computer systems designed to find complex patterns in data, which can be used to form rules. The system then performs tasks like speech recognition, decision-making and language translation.
Different industries are using AI to uplift business outputs through enhanced decision-making and operations.
AI in Agriculture
Australian farms are using AI for livestock monitoring, weed control, grain grading and classification, and more. With AI completing analysis on thousands and thousands of data points, farmers are alerted to issues much earlier, enabling a shift to proactive management.
Drones monitor livestock in real time from an office and ear tags track individual animal movements, health and biosecurity. The data generated from these innovations is then processed by AI systems, equipping farmers with early detection and improved the response time.
An intelligent irrigation system that automates water pump management in response to sunlight, temperature and humidity level inputs, without farmer intervention, has just won this year’s Sustainability category of the Financial Review AI Awards.
The system, developed by agtech company Aglantis and La Trobe University, is reported to deliver 20% water savings, up to 50% reduction in electricity costs and at least 30% labour savings.
Initially created to improve land management and prevent fertiliser run-offs into the Great Barrier Reef by reducing excess water applications, the data-driven automation takes guess work out of farming for sugar cane farmers in Far North Queensland.
AI In Mining
Rio Tinto’s AI-powered mining ecosystem combines driverless trucks and rail with predictive analysis. The Mine Automation System (MAS) operates like a network, pulling data from almost all sites and displaying it visually in a digital twin, called the Rio Tinto Visualisation. The data is also made accessible through the usual dashboards, graphs and charts.
AI systems are then used to analyse these vast amounts of data, using sophisticated algorithms, to optimise operations. One such outcome is a reduction in queuing, which is the time a truck spends idle while waiting for the next process. In some instances, AI has reduced truck wait times by 55% in open-pit mining.
AI is also used in Rio Tinto’s iron ore business to tackle an immense amount of data and determine optimal routes from mine site to customer. While the previous system could take up to a day, this new system maps routes in minutes, demonstrating AI’s applicability to transport and logistics.
AI in Transport and Logistics
Australian fleet companies are using AI to optimise routes, improve load efficiency and predict optimal maintenance on vehicles. Toll Group is using AI to process real-time sensor data across its 400-vehicle fleet in New South Wales, to undertake predictive maintenance.
Predictive maintenance uses data points across the truck fleet to monitor equipment against trained baselines. When an anomaly is detected, the system can prompt teams to open a maintenance ticket, in real time. Through predictive analysis, companies can also gain a better understanding of the root cause when breakdowns occur.
Through this system, Toll Group reports a reduction in unplanned downtime by 42% in the first 12 months. Maintenance costs were reduced by 28% and vehicle lifespan was improved by 15%. Processing huge amounts of data in real time enables smarter fleet management with tangible operational gains.
Hallucinations and human verification
AI usage is not without risk. As businesses work towards unlocking its efficiencies, it’s important that good governance and accountability guardrails progress as well.
AI-generated outputs, calculations and data can contain errors or even hallucinations, the most common of which are reference titles to articles or books that don’t exist. The AI system generates the pattern of a reference, a title, date, authors, without the reference actually existing.
This is why building human verification into any AI supported process or tool is critical.
Consulting firms have recently been in the news over AI-generated references used in reports. Late last year, one firm was required to provide a partial refund to the Department of Employment and Workplace Relations, after confirming AI hallucinations and errors were present in a report it delivered to them. Another firm’s report on AI usage included hallucinated citations, causing the report to be withdrawn globally.
These scenarios demonstrate the importance of double-checking AI generated outputs to minimise professional liability and reputational risks.
Reducing the risk
The Australian Prudential Regulation Authority (APRA) has issued a formal letter to industry about AI usage, noting its observations around increased security threats, governance, change management and supplier risk.
The Australian Securities and Investments Commission (ASIC) has also called for an urgent cyber uplift in response to AI accelerating cyber threats and lists several actions businesses can take to reduce risk.
Frontier AI models, the most advanced with leading edge capability, can expose online security vulnerabilities at unprecedented rates. Recently, frontier model Mythos 5 was removed from non-US nationals, following national security concerns.
Other practical steps business owners can take:
- Get familiar with how AI works, its strengths and its weaknesses
- Build good governance at all levels, from executives to employees
- Lock down the AI environment with an enterprise solution to protect IP and data
- Check the cyber health of your business
- Learn about current AI mitigations
- Review your cyber and management liability insurances.
Here to help
Introducing AI has the potential to bring about numerous benefits, but not without risk. Contact your EBM Account Manager to check that you are protected against AI risks.
