00:00King Charles has hosted several AI leaders in Scotland, where he urged them to keep artificial intelligence technologies under human
00:09control
00:09and ensure they serve people, communities and the environment.
00:16The meeting was held at Dumfries House in Ayrshire, home of the King's charity,
00:22and was attended by figures from many well-known AI giants, including OpenAI, Anthropic, Google DeepMind and NVIDIA.
00:33The King referenced the rapid speed at which AI has advanced over the years and continues to develop.
00:39He urged those in the room to find ways to use AI for beneficial purposes with safety at its heart
00:47and consider how to build international cooperation and consensus to achieve that.
00:53AI is already showing its immense capacity to improve and save life, for example in the field of life sciences
01:03and medicine.
01:05Yet those who have created these technologies are now increasingly warning that AI risks developing darker capacities.
01:16Perhaps even to take life.
01:19And in this regard, if I may say so, there seems urgency in adequately considering the existential dangers of such
01:28technologies falling into the wrong hands
01:31and being used in potentially catastrophic ways.
01:37Demis Hassabis is the chair and co-founder of Google DeepMind.
01:41He outlined some of the dangers involved with more advanced AI models.
01:46One is misalignment.
01:48This is where the AI system doesn't do what we instruct it to do or somehow circumvents its guardrails.
01:54And the second risk is misuse, where a bad actor repurposes the general technology that was built for good intent,
02:02but repurposes it for harmful ends.
02:04The chance of something bad accidentally happening, if we get things wrong, is definitely non-zero.
02:11So we should take these risks very seriously.
02:14The meeting comes in the wake of several high-profile incidents regarding the safety of AI technology.
02:20On Wednesday, OpenAI, which makes ChatGPT, reported six incidents of unexpected or concerning behavior by its models.
02:32NVIDIA CEO Jensen Huang, the company behind much of the infrastructure and computing components involved with AI,
02:40has criticized calls for slowing down the development of AI models.
02:46When a product is not safe enough, we should hold it back and keep engineering.
02:52We've always done that, and we should continue to do that.
02:56The ultimate purpose of AI is to expand access to knowledge and opportunity.
03:01Open models must be protected.
03:03Open models give researchers, universities, start-ups and countries a foundation they can adapt to their own languages, science and
03:13public services.
03:14While the UK does have a significant role in AI safety through its AI Security Institute and research community,
03:22the sector itself is dominated by competition between tech companies in the US and China.
Comments