• July 5, 2023
• StJohn Deakins
• no comments

At times, memes collide almost synchronously. 

Firstly, this month, the global premiere of the new Oppenheimer movie, an unlikely feel-good summer blockbuster, tells the story of the dawn of the Atomic Age and humanity’s horrific capability to obliterate entire cities, potentially extinguishing human life on Earth. 

Meanwhile, 2023 has seen the emergence of self-proclaimed ‘new Oppenheimers’. The developers of powerful new Generative AI Large Language Models (LLMs), such as Sam Altman at OpenAI, have repeatedly likened their new technology to the splitting of the atom, both in its potential economic value and its existential threat to humanity.

The parallel is not a coincidence. 

The ‘new nuke!’ narrative suggests that new legislation is needed to limit access to this new science and entrust its development to a small group of expert magicians. But is the ‘new nuke’ an accurate reflection of the potential of Generative AI for society? Amid the cacophony of exaggerated claims and apocalyptic warnings about AI, it’s crucial to separate fact from fiction and use analogies that help us progress towards better outcomes for all.

AI Futures

As we move towards a world of increasingly powerful and personal AI, the impact of this choice will become very personal to all of us. We are now firmly in the Digital Age and are undergoing a rapid Information Revolution. We have a couple of ways to view and treat AI, leading to two very different futures:

1) AI as the new scary nuclear weapons, threatening all of humanity;

2) AI as the new ubiquitous engines, upgrading human abilities and enhancing humanity.

 

The New Nuke

The ‘New Nuke’ narrative suggests that Generative AI marks an alarming acceleration towards a super-powerful and dangerous Artificial General Intelligence (AGI). Similar to nuclear weapons, this brand new technology is potentially an existential threat to all of humanity, the narrative being “It must be harnessed and contained!”. Therefore, as the thinking continues, we need to select a group of trusted experts to be the guardians ensuring the protection of humanity, while advancing the new secret ‘magic’. Hence, government-sanctioned and funded Command and Control mechanisms are required, just as they were with the Atomic Bomb. These could be lucrative even by Silicon Valley standards. The US Nuclear Programme spent $10 Trillion over a few decades.

The New Engine

Treating AI as the ‘New Engine’, on the other hand, suggests that AI is a new empowering utility, a powerful new tool.

Mechanical engines shaped our last era, the Industrial Revolution. Starting with steam engines, human muscle power was first replaced and then massively upgraded as the industrial era progressed. Today, we have combustion engines in cars, ubiquitous access to electricity, electric motors in washing machines, electric toothbrushes, mobile phones, jet engines, and rocket motors taking us into space. These all result from a small selection of the many types of engines that have become so normal and ubiquitous, such that they are now invisible to us.

While ‘New Engine!’ is a far less dramatic narrative than the arrival of ‘New Nuke!’, the arrival of the engine has had a far more profound impact on humanity. The engine led to industrialisation, urbanisation, electricity, mass transportation, Enlightenment thinking, political revolutions, emancipation of slaves and then women, the rebirth of democracy, the arrival of global trade, and universal human rights.

Personal Utility

Where mechanical engines have been force multipliers of muscle power, economies, and our global society, these new AI engines will be force multipliers for our brains and our cognitive power. As noted by Benedict Evans recently, automation will create new jobs and industries. AI Engines may also create entirely new economics, unlock new forms of politics, and proliferate throughout society. They will become useful in remarkable new ways that are completely unimaginable to us right now.

“We’ve grown an extra limb overnight that we don’t know how to use yet” – Charlie Booker, Black Mirror.

The technology is already with us and evolving quickly. A relatively simple Social Media algorithm likely nudged you towards this post. Netflix and Amazon have been perfecting their recommendation AIs for over a decade.

In the same week that OpenAI testified to the US Congress that their Large Language Model (LLM) AI is ‘too powerful’, Apple quietly announced LLMs in its iPhone for better auto-correct – without mentioning AI even once in the two-hour announcement. Apple is leading big tech in utilizing Generative AI as an engine. For them, it’s a new type of technology that completes real-life jobs in a better way. To maximize personal privacy and control, Apple is doing this locally on people’s own Apple devices, making the engines less scary and more easily acceptable in everyday life. Generative AI becomes simultaneously pervasive and invisible to us, like engines.

Personal AI

This points to where AI is headed. AI engines are likely to quietly work in all of the services we use. Ultimately, these service AIs will interact with our own AI agents – our very own personal AIs.

This is where our choice of futures becomes stark. If we relate to AIs as a new type of nuclear weapon, these deeply personal AIs will be provided by one or two huge and secretive companies. The ‘magic’ that powers them will be protected as secrets, including from you and me.

However, if we relate to AI as engines, there will be open competition and guardrails, making it more likely that personal AI is created transparently and controlled by you. Of course, to ensure this happens, there will need to be protections and regulations.

Safeguards

In the industrial era, engines arrived in waves. Each new wave of machines attracted new and dynamic legislation. From standards for early steam trains through to US federal legislation for car safety in the 1960s – as new uses for engines arose, regulations evolved dynamically to mitigate potential harms. We need legislation that governs these new AI engines.

Like any powerful new technology, there are, of course, very real risks. Far from these being distant possibilities, many of these issues are already with us, such as misinformation, biases, and deep fakes. Legislatures are kicking into gear. The approaches by both the European Union and the UK are a good start. 

The EU is setting a broad framework of proactive rules which sets a baseline and also mitigates against the ‘Nuke’ approach to AI. Meanwhile, the UK is using existing regulators and new bodies to create a cross-sector approach dealing with AI risks across society.  These are, of course, the first regulatory baby steps. Legislators will need to be ready to evolve laws as fast as the technology will undoubtedly progress. 

As we travel deeper into the realm of artificial intelligence, it becomes clear that the real power lies in harnessing AI’s potential for positive change, innovation, and advancement for all. A digital society where empowered digital citizens are served and uplifted by their own personal AI. 

CitizenMe is on this mission to empower everyone with better life outcomes using their own 360º personal life data. We achieve this by providing people with personal AI, working with their own data, all held and controlled on their own smartphones. This vision sees everyone with their own personal AI engines, which can be used collectively to create new economic and social value – and a sustainable digital future.