Defcon hacker convention11/9/2023 ![]() ![]() And while t he speeches are a highlight of the event, there are plenty of other unique and unusual things to do at Defcon. Participants must go through a rigorous review panel of their peers, before earning a coveted spot to speak at Defcon. And this year an unexpected participant was Tesla, apparently looking to recruit some computer geniuses. And one Rio hotel staffer told us that employees there were warned not to bring their cell phones, credit cards or anything else for the four-day conference because they could be breached.įor those looking for a Defcon souvenir that has not been hacked from someone, the chaotic vendor area has it all: artists hawking Defcon T-Shirts, security companies pitching the latest and greatest hacking equipment, how-to books and lockpicking kits. We were told to put our IDs, credit cards and other chip-embedded cards in some form of protective RFID shielding case while at the conference.Ī fast food restaurant nearby the hotel was hacked while we were there. These are hackers after all, and nothing electronic is safe from the prying eyes of their sniffer programs – bots whose only purpose is to latch onto your personal data. The Defcon conference, which is the ultimate assembly of paranoia, costs $220 in cash. These hackers mingle with National Security Agents (NSA) and Federal Bureau of Investigations (FBI) agents and other government guys who listen to the speeches, but really want to recruit talent. Some of the attendees include white hat hackers (good guys) and black hats (unscrupulous guys). Nearly 16,000 eager hackers of all ages showed up for Defcon 22 at the Rio Hotel. Thousands more who want to emulate his skills were in town for Defcon – a very different type of security conference that’s held the first week of August in Las Vegas and stands in stark contrast to the high-priced corporate event. Many of those acolytes descended on Las Vegas for a chance to see him speak at the Black Hat convention. The topic of Satellite vulnerabilities alone was enough to raise concern within the security industry and earn adoration within the hacking world. It costs about $3,000 to get in the door. We caught up with the Madrid-based IOActive security researcher right after he spoke at Black Hat, a high-end security conference geared toward corporations, security professionals and FBI types. As AI chatbots continue to evolve and become more integral to various industries, addressing their security vulnerabilities becomes crucial.Water safety stressed in wake of recent Lake Michigan drownings The major AI players have made commitments to prioritize security and safety, but smaller competitors may not have the necessary security measures in place. ![]() ![]() There are concerns that AI chatbots can be manipulated for financial gain and disinformation, with potential threats to privacy and company secrets. The authors of the book “Not with a Bug but with a Sticker” described the state of AI security for text-based and image-based models as “pitiable.” National Security Commission on Artificial Intelligence highlighted that attacks on commercial AI systems are already happening, and protecting AI systems has been an afterthought. The lack of security in AI models is a widespread issue. For example, one researcher tricked a Google AI system into labeling malware as harmless with just a simple line of text. Despite efforts to enhance security, researchers have already exposed vulnerabilities in AI systems. The competition at DefCon aims to uncover new security problems with AI chatbots, similar to the early days of computer security. They lack security measures as they were primarily focused on training data collection, resulting in racial and cultural biases. Academic and corporate research shows that AI models are currently too unwieldy, brittle, and malleable. However, it is expected that fixing flaws in AI chatbots will require significant time and financial resources.Ĭurrent AI models are vulnerable to flaws and biases due to their training process, and they can be easily manipulated. The findings of this competition will only be made public in February. Over 3,500 competitors have participated in this first-ever independent “red-teaming” of multiple models. At the DefCon hacker convention in Las Vegas, a three-day competition is being held to expose flaws in eight leading AI chatbots. ![]()
0 Comments
Leave a Reply.AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |