OpenAI’s Post

Zico Kolter, we wish you great wisdom and mind clarity in your board of directors role with #openai. With power comes responsibility, and we wish you success in this responsibility. #AI #AIfuture #futureofhumanity #Aiandus #futureofwok #futureofdignity #technology #aiapplications #aiandprogress #aiuse #aiinventions #governance #aigovernance #airegulation

On paper, the formation of #OpenAI's new Safety and Security Committee reads well -- though one has to wonder how objective a band of insiders can be. #AI #ArtificialIntelligence #ResponsibleAI #TechCrunch

In this episode, we're joined by Mandy Andress (CISO - Elastic) as she discusses how to securely integrate enterprise data with OpenAI and other LLMs. Mandy shares her perspective on the cautious approach to AI, privacy concerns, and the opportunities it presents in the realm of security and data analysis. She also delves into the ethical implications of AI innovation and the growing need for global conversations and a balanced approach to moving forward. Listen to the full episode ⤵ https://lnkd.in/g8E9DvFq

In this episode, we're joined by Mandy Andress (CISO - Elastic) as she discusses how to securely integrate enterprise data with OpenAI and other LLMs. Mandy shares her perspective on the cautious approach to AI, privacy concerns, and the opportunities it presents in the realm of security and data analysis. She also delves into the ethical implications of AI innovation and the growing need for global conversations and a balanced approach to moving forward. Listen to the full episode ⤵ https://lnkd.in/g8E9DvFq

Great video interview discussing AI and security. critical topic as we delve into this new technology. https://lnkd.in/gHqDuaq9

Interesting development, but not appropriate for safety development by itself. Individually blocking out troublesome topics is like whack -a-mole. I don't think they can even undertake a lobotomy equivalent excision, because topics are not clustered. In any case, this is nothing even close to Asimov's Three Laws. Do not use an LLM for anything important. Also interesting to note that the clowns at OpenAI have whacked their safety team, and then signed the second round of the Bletchley declaration. Hypocrites. https://lnkd.in/e2cYAxJ6

🌟 OpenAI Establishes Safety and Security Committee 🌟 In a significant move to enhance the safety and security of its AI projects, OpenAI has formed a dedicated Safety and Security Committee. This committee, led by directors Bret Taylor (Chair), Adam D’Angelo, Nicole Seligman, and Sam Altman (CEO), is tasked with making crucial recommendations to ensure the safety and security of OpenAI’s initiatives. 💡 Why This Matters: OpenAI is on the verge of training its next frontier model, which is expected to elevate the capabilities on the path to Artificial General Intelligence (AGI). With great power comes great responsibility, and OpenAI is committed to maintaining industry-leading standards in both capabilities and safety. 🛡️ Key Responsibilities: The Safety and Security Committee will evaluate and enhance OpenAI’s safety processes and safeguards over the next 90 days. Their findings and recommendations will be reviewed by the full Board, followed by a public update on the adopted measures. 👥 Expert Team: The committee includes top OpenAI technical and policy experts like Aleksander Madry (Head of Preparedness), Lilian Weng (Head of Safety Systems), John Schulman (Head of Alignment Science), Matt Knight (Head of Security), and Jakub Pachocki (Chief Scientist). They will also consult with external safety and security experts, including former cybersecurity officials Rob Joyce and John Carlin. 🔄 The Path Forward: This initiative underscores OpenAI’s commitment to robust debate and continuous improvement in safety protocols as they advance towards AGI. The proactive formation of this committee ensures that safety and security are prioritized at every step of their innovative journey. Stay tuned for updates on the committee’s recommendations and how OpenAI continues to set benchmarks in AI safety and security. #OpenAI #AISafety #TechInnovation #AI #Security #AGI

Breaking: Sam Altman Steps Down from OpenAI's Safety and Security Committee 🚨 1⃣ Committee to transform into an independent board oversight group chaired by Carnegie Mellon professor Zico Kolter, with members like Quora CEO Adam D'Angelo, retired General Paul Nakasone, and ex-Sony EVP Nicole Seligman. #OpenAI #TechNews #SafetyFirst 2⃣ The move follows a letter from U.S. senators questioning OpenAI's safety policies and high staff turnover focused on AI risks. Researchers accuse Altman of prioritizing corporate interests over AI safety. #AISafety #Policy #EthicsInAI 3⃣ Despite Altman's exit, the committee retains authority to delay model releases over safety issues. OpenAI is enhancing its launch processes, increasing federal lobbying, and Altman joins the DHS AI Safety Board. #TechPolicy #AIFuture #SecurityMatters Read more: https://lnkd.in/e6niDD_F

Exploitable by Default: Vulnerabilities in GPT-4 APIs & Superhuman Go AIs with Adam Gleave of FAR AI Tune in to this episode of Cognitive Revolution, as Nathan Labenz sits down with Adam to discuss testing AI models, open source's role in AI safety, vulnerabilities of superhuman Go & more. Key Highlights: 🚨 AI exploits: from accidental jailbreaking to malicious attacks.. 🛡️ Deep dive into GPT-4's exploits and security risks. ⚖️ The robustness tax. 🕵️♂️ Unveiling flaws in superhuman Go-playing AIs. 🌐 Empirical scaling laws for adversarial robustness 🎥 Watch the full episode at: https://lnkd.in/gUw5ATh5