Making AI Trustworthy at Scale: The Neurosymbolic Path Forward (Butcher, Tunstall-Pedoe) | DLD26 скачать в хорошем качестве

Making AI Trustworthy at Scale: The Neurosymbolic Path Forward (Butcher, Tunstall-Pedoe) | DLD26

William Tunstall-Pedoe, founder and CEO of UnlikelyAI, and journalist Mike Butcher (Pathfounders) discuss a groundbreaking shift in artificial intelligence: the pursuit of trustworthy, explainable, and highly accurate AI at scale. Tunstall-Pedoe introduces the concept of “neuro-symbolic AI,” a hybrid approach that combines the power of neural networks with the precision and reliability of traditional algorithmic software. Unlike large language models (LLMs) that rely on statistical guesses and require immense power, this approach aims to offer both the capabilities of cutting-edge AI and the transparency businesses and regulated industries urgently need, Tunstall-Pedoe explains. In many enterprises, AI pilot projects often fail to reach production because systems occasionally go wrong, can’t justify outputs, and don’t fit regulated requirements. That’s why UnlikelyAI focuses first on finance, insurance, accounting, and other regulated industries where a bad decision is costly, as you’ll hear. Watch the video for further insights into this innovative approach to making AI more trustworthy. Discover inspiring talks from past DLD Conferences and catch the highlights on the this channel. Stay updated with thought leaders, groundbreaking ideas, and key moments from our events! For news, upcoming events and more interesting topics make sure to also visit: DLD Website: https://dld-conference.com DLD on LinkedIn:   / dld-conference   DLD on Instagram:   / dldconference   DLD on Twitter:   / dldconference   DLD on Facebook:   / dldconference