Silicon Valley's AI warning

Back to Morning Sun 2, the Bay Area is at the center of the AI boom, and now local tech leaders are grappling with how to make sure that boom does not create new dangers. Anthropic CEO is warning about cyber attacks, economic disruption, and loss of control over AI systems. At the same time, OpenAI and Google are discussing a possible new safety organization. To expand on the pack to slow down AI, we welcome Sanjay Goel, Professor and Chair Information Security and Digital Forensics at the University of Albany. Good morning, Professor. Thank you for joining us. All right. Thank you, Pam. Thank you for getting me. We feel like I have to ask you, is this too little too late? I mean, the cat's already out of the bag with this AI. It feels like it's hard to slow down now at this point. Well, I completely agree with you, right? So, but the thing is that AI is also causing a lot of development. There'll be transformational changes coming from AI, the way we work, the way we live. Things will be a lot easier. But at the same time, there are existential threats, right? The main threat being that, you know, will we lose control of our humanity to AI, right? The super intelligence is coming through now. Already the models that we have are very strong. How strong will they get? Will they subsume all the work we are doing? Will it create huge employment opportunity problems? Right? So that's the kind of thing we're looking at. But at the same time, we will see a lot of rapid drug discovery, more scientific research. But there are benefits to that as well. So it's a mixed bag. We do need to somehow figure out what the existential threats are ahead of time so that we can manage the process better. Right. You mentioned so many things that can be helped by AI and some, you know, with medicine and research and things that can certainly benefit us. What is your biggest concern in trying to corral this technology and making sure that it's safe? Well, if you look at the AI models, which are super intelligent, they're being used in cyber warfare now, in real warfare. We have drones, which are AI based, which are automatically going and shooting. What prevents somebody from setting up a drone to basically identify somebody in the streets and keep roaming around and shoot them when they find them. Right. So there are a lot of malicious ways in which it can be used. Go ahead. No, no. Go ahead. The second worry I have is about the employment. Right. AI, let's say self-driving cars. How many people around the world are drivers right now? Literally in millions, closing to 500 million or something. What happens to their employment? A lot of entry level people, their jobs are going to be subsumed. It's not that AI will do all of them, but it will make people so much more efficient that one person can do the job of six people. Right. So that reduces employment. What happens to all of the people who are unemployed? Yeah. And talk about the, you know, you're talking about jobs, which is certainly a real concern. But, you know, we're talking about the safety concerns of it. I mean, last year, last week, we had the dire warning that AI has the capability of wiping out humans. You know, do you think that within 10 years? Do you think that is a possibility? And what needs to be done to make sure that doesn't happen? Who should be in charge of those guardrails? Well, two things. First of all, I don't believe that humanity will be wiped out within 10 years by AI, because as AI is coming through, we are also understanding more how we need to react to it, how we need to adapt and adjust to it. Right. So, yes, people can use AI. Who can use AI? We have an adversary, right, all over the world. And if they get hold of AI, they can create basically weapons which can hurt us. They can, AI can take control of people. But the most important thing is AI does not still have feelings. It's not sentient. So it does not have a feeling of its own that it wants to come and destroy humanity. It's people who will be using it to cause harm. And that's still true today. So I don't think that we will reach an extinction-like event, but will it change society? Will it create a disparity between the haves and the have-nots? Absolutely. I think as we have less and less need for people and people's work, and AI is taking more and more work, people who control AI are going to control the economy, are going to control the workforce. So I think that's a more social danger that we have from AI, more than just making the population extinct. Okay. And unfortunately, really quick before we let you go. Who do you think should be in charge of this? Should it be the government, government along with the tech CEOs, a collaborative group? Who do you think should be in charge of these guardrails and making sure it's safe? Right. I think we need to have general people involved in this. It's not in the CEOs who should be making decisions for everybody. I think the government should be involved, politicians should be involved, the CEOs should be involved, and general public experts should be involved in the whole thing. People who understand the ethical dilemmas here, it's all ethical questions, right? How much do we advance science? At the same time, how much do we restrict it to protect humanity? I think all of those questions are very philosophical, as well as scientific. So I think we need a combination of people with different expertise. All right. We're glad that someone like you is involved in this discussion as well. Sanjay Goel with the University of Albany. Thank you so much for joining us. Thank you very much, ma'am. I appreciate this.