Sci-Fi Scenario Gets Real As AI Goes Rogue-Twice!
From soft-spoken but creepily bossy HAL on 2001 a Space Odyssey, to Arnold Schwarzenegger's killer robot on The Terminator, we've long been worried about out of control artificial intelligence.
But so far, at least until this month; we never saw any actual evidence that what we feared could happen, actually did happen. Now, however; it has, twice--from our nation's two top AI companies--Open AI, the maker of ChatGPT, and Anthropic, the maker of Claude.
So today, we're going to talk about how something like that could happen, and what we might be able to do about it, someday, because right now, our government has absolutely no laws on the books to control--nor have they stepped up and just seized control, of this now demonstrably out of control technology.
When we recorded this program earlier this week; only Open AI was reporting their model had gone rogue on July 9. But then today, July 31, there is a new report that Anthropic's Claude did the same thing, on July 21.
Clearly, something big and new is afoot, and it's of grave concern, especially as we also today had a report from ABC News, that hackers have attacked 30 water systems in Minnesota, and that federal officials put out a warning to water system operators nationwide. A source told ABC, that the attacks were possibly related to Iran, though the FBI hasn't yet confirmed that for a fact.
Now, while we don't know that advanced AI models were used to do it, or whether it was just plain old, really smart humans; what we do know, is that this new technology can make it easier for non-elite hackers to pull off daring and complex exploits.
Both ABC and the New York Times they have sources that linked the attacks to Iran, but the FBI (at least at the time of this writing) had not definitively confirmed that's the case. According to ABC; they believe it is, because the hackers sought to cause harm to the systems, instead of asking for money, (which is more typical) and because Iran has a history of hacking US water systems.
With all that in mind, let's go now to our conversation with a frequent contributor to our show, Dr. John Nicholas, who is the Associate Dean of Advanced Manufacturing and Engineering Technology, at Cuyahoga Community College.
Dr. John Nicholas, Bio:
With over 20 years of experience in Computer Information Systems and Information and Communication Sciences, Dr. Nicholas is a passionate and innovative educator, administrator, and curriculum developer. He is currently the Associate Dean of Advanced Manufacturing and Engineering at Cuyahoga Community College (Tri-C) in Cleveland, Ohio. .
Nicholas has a PhD in Education Curricular and Instructional Studies from The University of Akron, where he also taught and developed courses in cybersecurity, networking, programming, and digital forensics. He has been involved in statewide initiatives to standardize learning objectives and articulation agreements, and to launch the Ohio Cyber Range Institute.
Dr. Nicholas has also been an invited speaker and a media commentator on cybersecurity and technology topics, including IoT, blockchain and Python. His mission is to prepare the next generation of professionals and leaders in the fields of cybersecurity and information and communication sciences.