Andrew Schroeder’s Post

📢 👇NEW conversation on humanitarian #AI and the Meta #Llama impact grants program with the always brilliant Andrew Schroeder. 🙏 Thanks to Brent O. Phillips for the opportunity and the excellent questions 👉 made us reflect deeply on the opportunities and challenges of leveraging #AI for decision support at times of crises! We also cover the necessary conditions that need to be in place for the sector to be "ready" for using #generativeAI and how services such as #Lama can help spur innovation in the space,...and much more (requires a whole series 😀 ) Check it out below. 👇

Stefaan Verhulst, PhD G. Verhulst and Andrew Schroeder speak with Brent O. Phillips Phillips about Meta's Llama Impact Grants program supporting applications of AI for social good on Humanitarian AI Today. We also discuss collective intelligence, large language models, data accessibility and making data AI ready, data collection and standardization initiatives geared for humanitarian actors, the impact of AI on humanitarian operations, localization, and how humanitarian actors can collaborate around advancing humanitarian AI #aiforgood #humanitarianai #humanitarian #artificialintelligence #machinelearning #podcast https://lnkd.in/gNRJHanu

Stefaan Verhulst, PhD and Andrew Schroeder speak with Brent O. Phillips Phillips about Meta's Llama Impact Grants program supporting applications of AI for social good on Humanitarian AI Today. We also discuss collective intelligence, large language models, data accessibility and making data AI ready, data collection and standardization initiatives geared for humanitarian actors, the impact of AI on humanitarian operations, localization, and how humanitarian actors can collaborate around advancing humanitarian AI #aiforgood #humanitarianai #humanitarian #artificialintelligence #machinelearning #podcast

🧠 Exploring the World of Hallucinations in LLMs (Large Language Models) 🤖 Hallucinations in AI, particularly in Large Language Models (LLMs), have been a topic of growing interest and concern. These intriguing phenomena raise important questions about the capabilities and limitations of artificial intelligence. 🤯 🌐 In a recent study, we delved into the fascinating world of LLM hallucinations. Our research sheds light on: 1️⃣ Understanding the Causes: What triggers hallucinations in LLMs? Is it the data they're trained on, their architecture, or something else entirely? 2️⃣ Ethical Implications: How do LLM hallucinations impact decision-making processes, misinformation, and biased content generation? 3️⃣ Mitigation Strategies: Can we develop methods to reduce or prevent hallucinations in LLMs, ensuring they provide reliable and accurate information? Join the conversation on the future of AI and its potential impact on society. Let's explore the boundaries of what's possible and what's responsible in the world of artificial intelligence. 💡🤝 #AI #LLMs #HallucinationsInAI #EthicsInAI #AIResearch

Join us this upcoming Wednesday the 5th of June at 3 pm Rome time for an EvalForward webinar during the Global Evaluation Initiative week. In this gLOCAL event, we will explore the transformative role of Artificial Intelligence (AI) in evaluation practices and its implications, drawing in particular on:  United Nations Population Fund (UNFPA)'s Latest Strategy: Explore UNFPA's latest strategy for responsibly leveraging AI in evaluation. Discover how to navigate ethical considerations surrounding AI use in evaluation, and gain insights into a roadmap focusing on specific evaluation use cases. World Food Programme's approach to building an Artificial Intelligence evidence mining tool: Learn about the ongoing journey of implementing a corporate AI tool, exploring the challenges and lessons learned along the way. Practical storytelling: Gain practical tips and insights to enhance accuracy and efficiency in key evaluation activities such as analysis, writing, and review. Through engaging storytelling, we'll showcase effective strategies to harness AI's potential in improving evaluation practices. You can register here: https://lnkd.in/dZ4E2faM

AI Experts Discuss Technology Set to Change History at Bowdoin #AIRevolution 🤝 Follow us on Discord 🔜: https://lnkd.in/gt823Zd3 _ ❇️ Summary: The users of AI are asked to trust that it will benefit them and that companies making AI have their best interests in mind. However, generative AI and its outputs are making people trust each other and the information they receive less. The panelists discuss the importance of trust in AI, the need for validation, and the potential risks of blindly relying on technology and companies. They also emphasize the need to define productivity beyond contributing to GDP and to reflect on the alignment of AI systems with human values. The humanities are seen as crucial in addressing the alignment problem, defining competence, control, and consciousness, and connecting AI to larger projects of enhancing freedom and dignity. It is emphasized that understanding the human condition, sociological drives, and historical antecedents is essential in avoiding the negative effects of technology. The panelists also highlight the incredible capabilities of language models in AI and how it has revealed the extent of human language abilities. Hashtags: #chatGPT #AIAdvancesChangingHistory #BowdoinTechPanel

WASH 💧 Talk Announcement: Tune in on Thursday, March 7th at 9:00 PT (12:00 ET) for the next talk with Olivier Mills, Baobab Tech Founder, discussing "WASH Knowledge and AI". In this talk, Olivier Mills examines how AI, now an integral part of everyday life, can also play a transformative role in the WASH sector. Focusing on knowledge transformation and capacity building, we will explore how AI, large language models, and machine translation models are forcing us to rethink how we can democratize knowledge. Olivier will also demo a tool under development called WASH AI. Register here: https://t.ly/yhSr_