A Modern Enigma

The stakes couldn't be higher when it comes to the future of artificial intelligence (AI). Will it be our saviour or lead us to our downfall? This question has prompted world leaders, tech moguls, and AI researchers to converge at the UK's Bletchley Park this coming week (1st & 2nd Nov) for the first global AI Safety Summit. The gathering aims to explore how to harness the immense power of AI while mitigating its risks.

Yet, what exactly are these "risks"? The summit focuses mainly on "frontier AI," or AI systems so advanced they don't yet exist. However, critics argue that this futuristic lens sidesteps more immediate concerns. For instance, AI's enormous energy consumption and its impact on jobs are urgent issues, overshadowed by discussions about potential bio-terrorism or AI systems going rogue. Should we be losing sleep over hypothetical doomsday scenarios while ignoring the tangible issues that are already making an impact? I think this one case where whatever you choose to focus on, it’ll be the wrong thing for some.

Rishi Sunak, the UK Prime Minister, has been adamant that there's no cause for alarm, stating an ambitious goal for the UK to become a global leader in AI safety. Yet, what does "safety" entail in a world where AI's scope is ever-expanding? One challenge facing the summit is its participant list. While high-level executives from US tech giants are confirmed attendees, the absence of other key players like US President Joe Biden (VP Kamala Harris is attending), Canadian Prime Minister Justin Trudeau and German Chancellor Olaf Scholz raises questions about the summit's efficacy. Moreover, China's invite to the summit has sparked controversy, given its strained relations with the West. But how can we talk about global AI safety without a truly global dialogue?

In addition to the government and academia, commercial enterprises have a pivotal role to play. Firms like British-based Stability AI are at the forefront of AI development and call the summit a "once-in-a-generation opportunity" for the UK. Diversity of thought is critical in a landscape where the rules are yet to be written, and there is no doubt that (to the extent that they even do) the tech companies understand this space far better than governments.

Interestingly, there are increasing calls for a global body to oversee AI, a sentiment echoed by Antonio Guterres, United Nations Secretary General, who is also attending the summit. But what form should this oversight take? Yoshua Bengio, one of the "Godfathers" of AI, suggests a registration and licensing regime for frontier AI models, a first step towards global governance. Yet, will small steps suffice in a world where the pace of AI advancement defies belief?

The summit's objectives are grand, but its focus may be short sighted. AI's broader societal impact needs to be part of the conversation— from the data used in training these algorithms, which can perpetuate bias and discrimination, to the imminent threat of AI-induced unemployment. What's more, a recent study revealed that AI's future energy consumption could rival that of an entire country, yet this topic won't be discussed at the summit.

The Bletchley Park summit is a landmark event, but does it risk becoming a missed opportunity by focusing too narrowly on extreme risks? As we contemplate the future of AI, we must ask ourselves: Are we aiming our spotlight in the right direction, or are we merely captivated by the shadows it casts?

If this subject interests you, then I highly recommend Mustafa Suleyman’s book “The Coming Wave” for a detailed overview of all that is happening in the AI world today and its implications for society at large. It is truly astonishing in it’s scope and written by somebody, as one of the founders of Deep Mind, who has been at the heart of AI for more than a decade. Check out the site below:

See you Wednesday for some more AI news, tools and prompts.