What AI policies does industry want from Truss’s government?
Now that the new UK government is in place, those in the tech industry are keeping a close eye on the policy direction of Prime Minister Liz Truss when parliament resumes – particularly when it comes to artificial intelligence (AI).
The UK is often ranked third globally for AI, behind the US and China. The government has been keen to ensure Britain maintains its standing and create an environment in which another DeepMind – bought by Google for £400m in 2014 – can launch and thrive.
The Department for Digital, Culture, Media and Sport (DCMS) under the leadership of former Prime Minister Boris Johnson, published the National AI Strategy in 2021, which outlined the importance of supporting the technology and those working in it.
But there is uncertainty over the direction this and other policies will take under a Truss government. Here’s what some members of the AI industry are looking for when it comes to AI policy under the Truss administration.
The debate surrounding the regulation of the AI industry is similar to many in tech and business: balancing safety and innovation.
“While some regulation can be seen as beneficial, it needs to be not so much so as to hinder the development of AI,” said Sarah Pearce, head of UK data privacy practice at the law firm Hunton Andrews Kurth.
The government has taken some steps with AI policy, but there is concern from some that uncertainty over a solid AI assurance framework is hampering innovation.
“The UK government needs to prioritise putting regulations in place,” Heather Dawe, head of data at UST told UKTN.
“Until we have such regulation in place, the AI industry will not begin to reach its full potential across industries. We also need such AI assurance frameworks to ensure AI in use across society is fair, ethical, safe, and fit for purpose.”
A key area that needs to be addressed by AI policy, according to Dawe, is diversity.
“The government also must make progress in increasing diversity amongst those who develop AI,” she said.
“A lack of diversity in AI developers translates into biases and limits AI effectiveness across society.”
Bias in AI has become a leading criticism of the technology. This is particularly seen in recruitment. The concern is that automated systems used by hiring managers may unfairly rule out candidates from diverse backgrounds due to bias being programmed into the system.
The Information Commissioner’s Office (ICO) launched an investigation into discriminatory AI recruitment in July.
Resources and infrastructure
Raza Habib, the CEO of AI startup Humanloop – which in July raised £2m – believes that the UK is in desperate need of investment into computing infrastructure for the country to keep pace globally.
“The UK [AI] industry is at a massive disadvantage to our American counterparts when it comes to compute access,” Habib said.
Habib told UKTN that the innovations being done in the sector require “specialist computers” that the UK simply doesn’t have enough of in comparison to the US and is not investing enough to change that.
“People are still underestimating the rate of progress. Investment in this area, even if it might seem large, is still very small relative to the potential return,” Habib added.
“The potential value to be released is orders of magnitude bigger than most people in society, and especially in government currently realise.”
The digital skills gap is seen by the AI community as an issue intimately tied up with the potential of the AI industry.
There is concern from the entire industry that the digital skills gap could stifle innovation and slow both the economy of the tech industry and the nation as a whole.
While the term digital skills can be applied to basic computing ability, for AI, the problem may be greater still.
“The number of people who have expertise in machine learning AI, deep learning at a deep level, it’s very concentrated in a small number of companies,” Habib said.
AI is a specialist subject that has the potential to attract billions, and yet the number of professionals capable of scaling the UK’s AI sector may be insufficient.
While the standard for tech education at the highest level in the UK is impressive, with the likes of “Oxford, Cambridge, and Edinburgh producing amazing machine learning graduates”, that can only result in so many AI specialists working in the country, said Habib.
For Habib, tech visas are an essential policy to ensure the country continues to have access to the best available tech talent. Tech visas were mentioned in the digital strategy in June and last month the government launched a scaleup visa for businesses to sponsor high-skill tech workers for a two-year stay.
However, the current iteration of the digital strategy has been criticised for unnecessary delays and overlooking freelance workers from less privileged backgrounds.