Daragh Murray’s Post

There Must be Some Misunderstanding by Gary Marcus There must be some kind of mistake – Phil Collins / Genesis Great example last night from Denis O. on #LinkedIn, that connects with pretty much everything have been writing about, from elephants on beaches to the lack of understanding in #largelanguagemodels. Examples like these highlight pretty much everything I have been saying here, from the very first post (The New Science of Alt Intelligence https://lnkd.in/e9c7CiHA ) to the last (https://lnkd.in/ef2YePHG). ➡️These systems don’t deeply understand their inputs, contra Dr. Hinton, and as pointed out here regularly (https://lnkd.in/ekJCMeNt). ➡️ #Compositionality (https://lnkd.in/eWTNmZgK) remains a problem. So does #negation. ➡️ #Reliability (https://lnkd.in/eXr_WXyi) remain a problem. ➡️ You shouldn’t count on these systems (https://lnkd.in/eScYayW7). ➡️ Scale isn’t everything, and hasn’t and won’t be enough to solve any of the problems I just listed above. ➡️ We really should consider trying some other approaches to #AI (http://rebooting.ai/) https://lnkd.in/eGWF97W8

WorldQuant Founder, Chairman and CEO Igor Tulchinsky shares his insights on the interconnection of human behavior and advanced technologies in his latest book, The Age of Prediction, with Christopher Mason.

🚀 Excited to share my first blog on Medium! 🎉 Dive into the very basics of Artificial Intelligence and explore how it's shaping our world. Perfect for anyone curious to know about what AI is. Check it out and let me know your thoughts! #AI #ArtificialIntelligence #BlogPost #BeginnerFriendly

🚨 Navigating the World of Deep Fakes: A Double-Edged Sword 🚨 In today's digital age, the term "deep fakes" is increasingly becoming part of our lexicon. But what are deep fakes, and why are they stirring conversations across the globe? Let's delve into this phenomenon in simple terms. Imagine a video or audio recording that looks and sounds exactly like a real person saying or doing something they never actually did. That's the essence of a deep fake - a product of sophisticated AI and machine learning technologies that can create convincingly realistic forgeries. The Impact of Deep Fakes on Society: 1. 💡 Innovation vs. Misinformation: On one hand, deep fakes showcase the remarkable advancements in AI, offering potential in creative industries and beyond. On the other, they pose significant risks by spreading misinformation, challenging our ability to discern truth from fabrication. 2. 🔍 Threat to Trust: Deep fakes can undermine trust in media, institutions, and even interpersonal relationships by making it harder to trust the authenticity of visual and audio content. 3.🛡️ Security Concerns: They can be used maliciously to manipulate stock markets, sway elections, and even incite violence by spreading fake news and propaganda. 4. 👩⚖️ Legal and Ethical Dilemmas: Deep fakes raise profound questions about consent, privacy, and the legal implications of using someone's likeness without their permission. 5. 🌐 A Call for Responsibility: The rise of deep fakes emphasizes the need for ethical guidelines, robust detection tools, and awareness to mitigate their harmful impacts. Deep fakes are a testament to AI's growing capabilities, reflecting the technology's potential to both innovate and deceive. As we stand at this crossroads, it's crucial to foster a balanced discourse on leveraging this technology responsibly while safeguarding society against its darker implications. In this digital era, understanding and addressing the implications of deep fakes is more than a technological challenge—it's a societal imperative. #DeepFakes #AI #DigitalEthics #SocietalImpact

Being human involves complex social dynamics and decision-making — a combination of elements that may shift with the evolving landscape of data analytics and AI.

Harnessing the power of AI to provide a comprehensive and accurate picture of global conflict. Fenix Insight monitors and maps global and local news, social media and the dark net 24/7 in multiple languages to keep users informed of the true picture on the ground right now. >> https://bit.ly/3PCPmPJ

Saw these videos on a flight, worth sharing: Bill Gates explains what we can learn from firefighter teams and proactively plan against outbreaks: https://lnkd.in/g6ix25EH Greg Brockman explains about the inside story of ChatGPT and GenAI, and how it is about "Humans and AI" and not "Humans vs AI": https://lnkd.in/gqv5YM7x Yejin Choi explains in same conference as above 😅 about How AI is incredibly smart yet shockingly stupid. How "we cannot build a rocket to moon, if we build a building taller an inch every year". We have quantity of data and compute, but we need the equivalent quality and unbiased data: https://lnkd.in/gpx29rd2

From the first generation https://lnkd.in/endi57BK This text reflects on how the current generation is unique in leaving behind digital traces of their lives, largely through social media platforms like Facebook and MySpace, in exchange for fleeting attention. Unlike previous generations, whose lives were less documented, our digital footprints are vast but may hold little lasting value. This accumulation of data—our posts, likes, and shared moments—might become a monumental but superficial archive, potentially of interest to future historians or AI systems aiming to simulate or understand past human lives. The piece questions the future relevance of these digital traces and whether they will truly enhance historical understanding or simply overwhelm with their volume, marking a shift from traditional forms of memory and record-keeping to a new, digital narrative of human existence.

Watching media minimize the atrocities recorded by on-the-ground sources in Palestine is immensely frustrating. Excited to share https://NewsCord.org, a partner project of T4P that'sdesigned to surface biases in news reporting through the application of generative AI. Its founders converged around the idea at a November Hackathon for Gaza, but its impact has far greater implications than just the current crisis. NewsCord has the potential to change the way we interact with the news. Their model aggregates and summarizes hundreds of stories, identifying common threads and alerting readers to specific inconsistencies, omissions, contrasting language choices, and other subtle narrative functions that many of us take at face value when we consume the news. The goal is a new generation of discerning news consumers who have the whole picture. Test drive NewsCord via the website or download it on Chrome: https://newscord.org

Last-minute reminder for anyone who's been wanting to join a Sea Change Together session but not yet made it... If you're free between noon til 2pm (UK time) today, Wed 24th, you're extremely welcome to join us. You don't even need to come for the full session if 2 hours is a bit much (I'm just there for that long, guaranteed, and you can dip in and out as suits). You can have your camera on or off, get involved with the discussion bit (from 12.10-ish today, we'll have Birgit Obermueller introducing our topic of the environmental impact of AI), or just have your mic off throughout and use the space to crack on with anything you need to work on! I'll pop the latest email (with Google Meet link) in the comments, as well as the link to sign up to emails for future sessions. You know, if you're the kind of person who needs more than an 8-minute heads-up – I'm told these people exist 😉