EU's AI agreement set to be watershed moment for regulation: Konrad Pietka

The proposed AI technology restrictions are set to be the strictest in the world.

They include the prohibition of biometric categorisation via the use of sensitive characteristics, such as political or religious beliefs, outlaw the exploitation of vulnerabilities, like age and disability, and also ban social scoring (measuring individuals based on how upstanding they are).

Additionally, a two-tiered approach is to be adopted, with stronger restrictions on more impactful AI models. The repercussions for non-complaint companies are severe, with fines ranging from €35m to 7 per cent of global turnover.

In the eyes of the EU, the outlined measures are meant to safeguard its citizens from any potential AI risks while simultaneously not imposing an “excessive burden” on firms in order to avoid stifling growth and innovation.

Despite this, the industry has not warmly received the news, with DigitalEurope, the continent’s technology association, perceiving the regulation as being overly restrictive.

Similar concerns have been reportedly voiced by European corporate giants Airbus and Siemens.

The deal is set to become law in 2025; thus, only time will show the legislation’s impact on AI development in Europe.

In the UK, recent data from the Office of National Statistics (ONS) has revealed that productivity has hardly grown since the 2008 financial crisis.

This phenomenon, coined the UK productivity puzzle, will likely serve as the main headwind to Chancellor, Jeremy Hunt’s ambition of increasing UK investments and growth.

ONS figures show that total factor productivity, a measure of resource allocation efficiency in the economy, was only up by an underwhelming 1.7 per cent between 2007 and 2022. In comparison, the productivity growth rate between 1992 and 2007 stood at 27 per cent .

Had this trajectory continued, UK productivity today would be over 20 per cent higher than in 2007.

The Chancellor seeks to rectify this problem by enacting “supply side reforms,” such as capital expenditure tax deductions, in a bid to improve efficiency across the economy; however, any measures would be limited in scale due to a historically high budget deficit and high borrowing costs.

Interestingly, the productivity problem is not uniform across sectors, with differing rates of innovation and technology implementation leading to an array of results.

For instance, productivity in the information and communication sector more than doubled since 2007, while the hospitality and retail sectors recorded a regress in the same period.

Moreover, productivity growth is being led by an ever-smaller cohort of firms, leading to a rise in market power and slow d own i n b u s iness dynamism. Therefore, to increase productivity, and in tu r n grow th and investments, these auxiliary issues would also need to be confronted.

Konrad Pietka is part of the Investment Research Team at Redmayne Bentley