Artificial Intelligence (AI) is increasingly becoming embedded in organisations, with over 90% of business leaders reporting its implementation. Its applications are vast: in healthcare, AI automates medical image analysis, aids in diagnosis, and predicts patient outcomes. In manufacturing, it enhances production efficiency by automating assembly and defect detection. Social and news media benefit from AI's ability to generate news feeds and articles quickly. The finance industry relies on AI algorithms for fraud detection, credit risk assessment, and automated investment strategies.
With opportunity though, comes risk. We’re increasingly seeing attackers exploiting AI vulnerabilities and it’s therefore never been more important that businesses implement AI security frameworks to combat this. Here’s what business leaders need to know.
AI can be a risky business
As AI develops and organisations increasingly rely on this tech, malicious incidents by attackers are becoming more prevalent. Tactics, Techniques, and Procedures (TTPs) used by attackers include exploits using facial recognition, privacy leaks and the generation of deep fake images and disinformation for political gain. Industry repositories like the AI, Algorithmic, and Automation Incidents and Controversies Organisation (AIAAC) are trying to crack down on these but often, tactics are sophisticated.
That’s where AI security comes in. AI security refers to the technical and governance considerations that harden AI systems to adversarial exploits. However, its lack of awareness is an issue. Only 14% of companies report awareness of consideration of AI security. By neglecting AI security though, businesses risk exposing sensitive information, falling victim to fraudulent activities and significant brand and reputational damage.
Making AI safe for businesses
AI is here to stay. What was once just researched in academia is now an important part of many business's lives and take up of this tech is only set to skyrocket. There’s an increasing parallel between the rise of the internet and subsequent cybersecurity threats, and the rise and adoption of AI technologies now.
Thankfully, there’s a growing number of solutions available to ensure businesses protect themselves against attacks, including our work at Mileva Security Labs. We’re an Australian-based start-up that launched in May this year, advising enterprises on safe, secure and responsible AI. I co-founded it with my colleague Dr Julie Banfield following my PhD researching AI Security at UNSW, and together we have a combined twenty-five years of experience in data science across consulting, academia, start-ups, and government.
Our participation in the UNSW Founder's New Wave incubation program in March 2023 paired us with the right professional network and industry experts to help us up-skill and launch our own AI security startup.
This led to the great honour of winning first place in the Doone Roisin Business Innovation Award at UNSW Founder’s New Wave pitch competition night. Today, Mileva Security Labs partners with companies to implement comprehensive AI Security Frameworks. Our approach not only helps businesses understand and mitigate the risk profile of their AI systems but also safeguards their customers from potential attacks.
The future of AI looks regulated
An open letter penned by the Future of Life Institute in March 2023, including signatories like Elon Musk, called for a pause in AI development. But I don’t think pausing is the answer.
Instead, regulation will be key. Australian businesses will soon need to comply with a raft of policies, governance and technical controls on their AI systems. Earlier this month, the Minister for Industry and Science, Ed Husic, released a discussion paper detailing exactly that. It’s set against a landscape where recent senate hearings in the United States (US) and meetings with tech giants like OpenAI, Google, Meta, and DeepMind underscore the political urgency of addressing AI security concerns.
In a future where AI use is only set to grow, AI governance, security and risk management play a vital role in building a secure environment for AI innovation. Mileva Security Labs is excited to partner with more businesses to ensure AI security doesn’t follow cyber security in becoming both a technical and geo-strategic threat and to ensure future generations can look forward to a safe, secure and prosperous AI-driven future.