This recent report from the thinktank Onward warns what differentiates this wave of automation and generative AI from previous ones in human history, which is its ability to take on creative, cognitive tasks, from writing to photography and graphic design. Once upon a time, humans could be persuaded that getting machines to do the drudge work would free them up for more interesting tasks. But AI is coming now for the dream jobs: well-paid, absorbing work done by people who love what they do and won’t let go easily. It’s coming not just for our ability to pay the rent, but for the things that make us happy.
Artificial Intelligence will be the transformational technology of our generation. In recent months the pace of change has accelerated, with the phenomenon of generative AI creating new applications that have exploded out of Silicon Valley and into the UK national consciousness.
A song that used AI to clone the voices of Drake and The Weeknd secured 15 million TikTok views before being hurriedly pulled from platforms. OpenAI’s GPT4, a Large Language Model (LLM), has performed better than 90% of American high schoolers across exams in nearly every subject. Google and Stanford researchers created a “mini-Westworld” where characters demonstrated “believable individual and emergent social behaviours”. Training compute has grown by a factor of 10 billion since 2010, and in the first three months of this year $11 billion has been invested in reaching Artificial General Intelligence (AGI), more than the previous ten years combined. The exponential rate of progress is therefore set to continue: the Generative AI revolution has only just begun.
While AI has landed squarely in the public debate through ChatGPT, the implications for policy and politics have yet to be fully recognised. AI presents an enormous economic and geostrategic opportunity
A recent Goldman Sachs report predicts that over the next ten years, the productivity benefits of generative AI alone will create $1.5 trillion of value. The scale of benefits in the UK will be driven by two trends:
Rapid adoption: ChatGPT garnered 100 million users in two months, compared to 30 months for Instagram, 55 for Spotify, and 70 for Uber.
Boosting worker productivity: A recent MIT paper found using LLMs reduced the time taken for professional writing tasks by 40% while improving quality. This productivity increase will also be felt in more advanced fields where the UK has a comparative advantage. DeepMind’s AlphaFold predicted the 3D structure of almost every known protein, a task that was thought to take decades of human labour
Given the potential of AI, international governments are already placing AI leadership at the centre of their geopolitical strategies. The UK’s Integrated Review in 2021 stated that “in the years ahead, the countries which establish a leading role in critical and emerging technologies will be at the forefront of global leadership”. But other countries are pulling ahead - the vast majority of recent AI advancements in LLMs have been driven by a few Silicon Valley-based AI labs with employees in the hundreds. The USA and China are gearing up for a technological battle for AI supremacy and the EU is seeking to lead on AI regulation. The UK has an opportunity to carve out a niche, shaping global standards and leading multilateral initiatives on AI safety. Google DeepMind, based in London, gives the UK a seat at the table.
AI will cause shocks to our labour market
Every wave of technological progress has generated fears, often unfounded, of rising unemployment. But this time the warnings should be heeded. For the first time technology looks set to automate cognitive functions and creativity, turning the traditional model of automating routine tasks on its head.
AI could cause superstar effects on steroids: Taylor Swift might profit from millions of personalised songs being created by AIs using her voice, or a high flying corporate lawyer could scale her productivity with effectively unlimited interns at her fingertips. White collar jobs - like the paralegals or graphic designers who make up a disproportionate amount of the UK workforce, particularly in London - could be automated partly or wholly out of existence. Research by OpenAI estimates that 19% of workers will have at least 50% of their tasks impacted while Goldman Sachs forecast 300 million jobs will be exposed. If the technology reaches AGI, where AI systems can perform any intellectual task that humans can, these numbers will multiply and the value of human labour will be limited.
The UK may also face macroeconomic shocks. Just as China’s introduction to the global trading system in the 1990s brought downward pressure on inflation, generative AI will bring down cost curves and prices for services. The structure of the LLM market and where economic value is reaped, including whether models become commoditised, will determine the impact on price levels and the necessary policy response. Downward pressure on prices may bring down short term interest rates, but much higher longer term interest rates may be needed if we move towards AGI. Markets are not currently pricing this in but the latest Metaculus forecast suggests AGI will be developed by 2032.
AI will create risks for our shared future
A recent letter signed by Elon Musk and Stuart Russell called for a six month pause in AI developments to address the issue of “alignment” – ensuring that AI systems act in accordance with human values and goals. In a recent survey of AI experts the median respondent estimated a 5% chance of AI posing an existential risk to humanity. The public is in agreement - polling by the Centre for the Governance of AI showed the majority of the public believe AI requires careful management. This is a global problem and coordination issues mean that multilateral solutions are essential - UK leadership will be required.