Try MagicTask: Your AI-Powered Task Assistant!

What's the future for generative AI? - The Turing Lectures with Mike Wooldridge

By - The Royal Institution
Overview
Certification
Reminders

Description

AI can now generate human-like language and artwork - but what other doors might it open in future? And how can we harness AI to make great leaps in technology possible? This talk was filmed at the Ri on 6 December 2023, in partnership with The Alan Turing Institute. Join this channel to get access to perks: https://www.youtube.com/channel/UCYeF244yNGuFefuFKqxIAXw/join Watch the Q&A with Mike here, exclusively for members: https://youtu.be/KSuUfU6x8rg Join 2023 CHRISTMAS LECTURER Michael Wooldridge for a fascinating discussion on the possibilities and challenges of generative AI models, and their potential impact on future societies. Find out more about the CHRISTMAS LECTURES here: https://www.rigb.org/christmas-lectures 00:00 What is machine learning? 05:54 How do neural networks work? 12:49 How Silicon Valley money created Big AI 14:50 The birth of Transformer Architecture 16:06 How was GPT-3 trained and created? 22:12 A massive step change in AI 25:45 How GPT-3 passed the 90s AI reasoning test 28:40 How has AI learned things it wasn't taught? 31:07 Chat GPT and how NOT to use it 32:26 Why do LLMs get things wrong so often? 35:29 The problems of bias and toxicity 39:12 Copyright issues with LLMs 42:23 Interpolation vs Extrapolation 45:44 Is this the dawn of General AI? 49:19 The different varieties of General AI 54:01 What actually is human general intelligence? 56:07 Is machine consciousness possible? ------ In partnership with The Alan Turing Institute we've been exploring the various angles of large-language models and generative AI in the public eye. Across three lectures, we aim to provide a comprehensive, thoughtful and engaging understanding of this rapidly emerging field and its impact on society. Watch the first lecture here: https://youtu.be/_6R7Ym6Vy_I And the second lecture here: https://youtu.be/s3_P22qf0IM ----- Michael Wooldridge is a Professor of Computer Science at the University of Oxford. He has been an AI researcher for more than 30 years, and has published more than 400 scientific articles on the subject. He is a Fellow of the Association for Computing Machinery (ACM), a Fellow of the Association for the Advancement of AI (AAAI), and a Fellow of the European Association for AI (EurAI). ----- Discourses are one of the Ri’s oldest and most prestigious series of talks. Since 1825, audiences in the theatre have witnessed countless mind-expanding moments, including the first public liquefaction of air by James Dewar, the announcement of the electron by JJ Thomson and over 100 lectures by Michael Faraday. In more recent times, we have had Nobel laureates, Fields medal winners, scientists, authors and artists – all from the cutting-edge of their field. Discourses are an opportunity for the best and brightest to share their work with the world. Steeped in nearly two centuries of tradition, a Discourse is more than just a lecture. The Discourse lasts exactly an hour, and a bell is rung to mark the beginning and end. To keep the focus on the topic, presenters begin sharply at 7:30pm without introduction and we lock the speaker into a room ten minutes ahead of the start (legend has it that a speaker once tried to escape!). Some of our guests and speakers dress smartly for our Discourse events to add to this sense of occasion. Read more about Discourses here: https://www.rigb.org/explore-science/explore/blog/history-friday-evening-discourse ----- The Ri is on Twitter: http://twitter.com/ri_science and Facebook: http://www.facebook.com/royalinstitution and TikTok: https://www.tiktok.com/@ri_science Listen to the Ri podcast: https://podcasters.spotify.com/pod/show/ri-science-podcast Donate to the RI and help us bring you more lectures: https://www.rigb.org/support-us/donate-ri Our editorial policy: https://www.rigb.org/editing-ri-talks-and-moderating-comments Subscribe for the latest science videos: http://bit.ly/RiNewsletter Product links on this page may be affiliate links which means it won't cost you any extra but we may earn a small commission if you decide to purchase through the link.

Course Content

What is machine learning?
1
5 min

What is machine learning?

5 Questions Ready
How do neural networks work?
2
6 min

How do neural networks work?

5 Questions Locked
How Silicon Valley money created Big AI
3
2 min

How Silicon Valley money created Big AI

5 Questions Locked
The birth of Transformer Architecture
4
1 min

The birth of Transformer Architecture

5 Questions Locked
How was GPT-3 trained and created?
5
6 min

How was GPT-3 trained and created?

5 Questions Locked
A massive step change in AI
6
3 min

A massive step change in AI

5 Questions Locked
How GPT-3 passed the 90s AI reasoning test
7
2 min

How GPT-3 passed the 90s AI reasoning test

5 Questions Locked
How has AI learned things it wasn't taught?
8
2 min

How has AI learned things it wasn't taught?

5 Questions Locked
Chat GPT and how NOT to use it
9
1 min

Chat GPT and how NOT to use it

5 Questions Locked
Why do LLMs get things wrong so often?
10
3 min

Why do LLMs get things wrong so often?

5 Questions Locked
The problems of bias and toxicity
11
3 min

The problems of bias and toxicity

5 Questions Locked
Copyright issues with LLMs
12
3 min

Copyright issues with LLMs

5 Questions Locked
Interpolation vs Extrapolation
13
3 min

Interpolation vs Extrapolation

5 Questions Locked
Is this the dawn of General AI?
14
3 min

Is this the dawn of General AI?

5 Questions Locked
The different varieties of General AI
15
4 min

The different varieties of General AI

5 Questions Locked
What actually is human general intelligence?
16
2 min

What actually is human general intelligence?

5 Questions Locked
Is machine consciousness possible?
17
4 min

Is machine consciousness possible?

5 Questions Locked
AI Mentor