AI Can Become Lethal if Not Controlled: Toby WalshUNSW Laureate Fellow Talks About Benefits and Pitfalls of AI
Toby Walsh, one of the most prolific AI researchers and laureate fellow at University of New South Wales, UNSW, is cautious about the short-term impact of AI on society. His book "Machines Behaving Badly: The Morality of AI" substantiates his research. One of the most remarkable innovations of AI is seen in the form of generative AI, which gives us synthetic text, video and audio. Notwithstanding its several benefits, bad actors have been using AI to generate deepfakes to rob banks, cause social upheavals, civilian coup subversion of democracies and cyberattacks. AI could soon become the tool of choice for hackers to orchestrate hyper-personalized cyberattacks.
"AI is not magic. Similar to other technologies, AI also raises grave concerns around ethics and morality," Walsh says. "However, unlike other modern technologies, AI is exceptionally autonomous. And without human oversight, it can become lethal and devastating."
In an exclusive video interview with Information Security Media Group, Walsh discusses:
- The ethical and moral concerns of AI;
- How biases creep into AI algorithms and act in a number of harmful ways;
- Pros and cons of generative AI tools like ChatGPT/GPT3.
Walsh is the chief scientist at UNSW's new AI Institute. He is also a laureate fellow and scientia professor of AI in the school of computer science and engineering at the university. Walsh has played a leading role at the UN to ban lethal autonomous weapons (aka "killer robots"). His advocacy in this area has led to him being "banned indefinitely" from Russia.