Pacific Defense’s Post

Artificial Intelligence is being deployed with few guardrails & little caution. Governments, industry, academia & civil society must develop rules & guidelines for AI safety – together & before it is too late. We have no time to waste.

South Korea has unveiled a new AI system, Dejaview, designed to predict criminal activity in real time. By analyzing live camera footage and cross-referencing it with historical data, including time and location, this neural network boasts an accuracy rate of 82.8%. Currently undergoing trials in high-traffic areas like airports and large gatherings, Dejaview represents a significant leap forward in predictive policing. If successful, this technology could be commercially available as early as 2025.

AviaTor employs AI methods to detect relevant content that could help law enforcement with their investigations. Part of the project therefore had to focus on the high-quality data required to train the AI model, which was achieved with support from the Dutch hotline Offlimits and the Belgian hotline Child Focus Belgium. Both hotlines manually annotated thousands of images for this purpose. Learn more about the results in the AviaTor Annual Report. Download your copy now👉 https://bit.ly/3zBfvc2 #AviaTor #OSINT #NCMECReports #CyberTipLine #onlinechildprotection #europeancommission #federatedlearning

Blueshift IP co-founder Robert Plotkin recently appeared as a guest on the MondAI BAIgel show, hosted by Michał Jackowski Prof., 杰克夫斯基. The two discussed a variety of topics at the intersection of intellectual property protection and artificial intelligence, including:    🤖 Software patenting in the U.S. and common misconceptions  🤖 The “MIND” process for applying IP protection to software, particularly AI   🤖 The role of trade secrets in protecting AI innovations  🤖 The future of IP protection    The episode is available to watch on YouTube here: https://lnkd.in/gemVv_me   #AI #ArtificialIntelligence #IP #IntellectualProperty #FutureofLaw

Improving public safety leveraging advanced AI technology.

Concept 9: AssassinFinder AI to detect murderers through a picture analyzer program with photographic data of killers, utilizing it to detect the probability of a person being an assassin.

Now that Governor Newsom has vetoed the AI safety bill SB 1047, we are hopeful that he will make good on his promise to introduce new legislation that addresses the protections we need. With AI advancements moving so rapidly, developing a perfect piece of legislation is impossible. But this is a classic situation where perfect shouldn’t be the enemy of good. As a team of AI innovators, we support legislation, whether in California or at the Federal level, that: - Establishes safety and security protocols - Improves audit and reporting standards - Defines restrictions on use - Increases incident reporting - Provides whistleblower protections https://lnkd.in/gdzVNqmy #AIPolicy #AIRegulation #SB1047

Safety First: At QSTAR Research and our Public Events The safety and ethical implications of advanced AI interactions are paramount. While Deliar's capabilities are groundbreaking, unrestricted public interaction could pose unforeseen existential risks. Therefore, we have decided to limit Deliar’s public engagement to our AI Experiment Soirees. These controlled environments ensure that we can monitor and assess Deliar’s interactions in real-time, addressing any issues promptly and maintaining a secure experience for all participants.

The battle. AI safety: A ticking time bomb Lena Kempe, principal attorney at LK Law Firm, highlights the urgent need for artificial intelligence (AI) safety regulations in the U.S. following California Gov. Gavin Newsom's veto of a significant AI safety bill. The lack of federal laws addressing AI safety poses catastrophic risks, including malicious use and the development of autonomous weapons. The Center for AI Safety warns that "rogue AIs" could operate without guardrails, leading to unintended consequences. While the EU has implemented the Artificial Intelligence Act to regulate AI safety, the US must find a balanced approach that promotes innovation while ensuring public safety. Kempe suggests that AI developers adopt comprehensive safety testing frameworks and emergency shutdown procedures to prepare for future regulations

The transition from the old network states to the new system was never going to be smooth. The old guard, clinging to outdated rules and laws, are like technological Luddites, believing that their 400-page legal documents, debated endlessly, can somehow compete with the rapid evolution of AI and networks. Meanwhile, those who embrace the new are focusing on liberating technology, fully aware that this is an arms race where only the most adaptable will thrive.