Artificial Intelligence At The House Of Commons
Andrew Burgess – one of D/SRUPTION’s expert contributors on Artificial Intelligence – provided evidence to the All Party Parliamentary Group on AI
This article is an adapted version of his talk…
In my day job I’m a strategic adviser on Artificial Intelligence, which means that I help organisations and enterprises develop their AI strategies so that they can best exploit these new and disruptive technologies. I also advise technology companies and consultancies, from startups to the Big Four, on how to embed AI into their service offerings.
I’m explaining this detail about my experience because, I want to make the point that I am working at the ‘coal face’ of AI in business: I’m focused on how AI can help businesses today.
AI is a technology that can potentially benefit all, and therefore it should ideally be available easily and without restrictions to as many people as possible. But, just as AI can potentially wreak massive damage at scale, we know that it is subject to many constraints, whether they be around the specialist knowledge required, the issues around data flows across borders or the rights to data privacy for the users of AI.
My vision to ease these constraints is for a democratised AI. This is where AI is being used every day by citizens, consumers and businesses in an open, transparent way. It’s where access to the knowledge to create AI is easily available, and where AI has moved beyond the hype-and-fear bubble and is focused on delivering real value to society rather than just click throughs on adverts. Crucially, an AI that is being implemented and used by the people will be for the good of the people, with the necessary controls and balances inherently built in.
In my work I see some great things being done by technology companies to help democratise AI. It’s good news that much of the leg work of data scientists can now be done on data science platforms – in effect using AI to find the best AI algorithm. But this is still a long way from where we need to be. For example, I know that it is easier and better to get data scientists from Estonia or Portugal than it is from the UK. I know that the vast majority of business people I speak to have no real idea what AI is and how it can help their businesses. Through no fault of their own, these people get distracted and confused by the headline stories of the big AI companies touting their alleged successes.
Many people are advocating for global standards in AI to try and level the playing field and protect users’ rights. As I discussed recently in D/SRUPTION, I firmly believe that it is a fool’s task to try and create global standards for AI, particularly around ethics and trade. There are two huge hurdles to overcome. The first is the diverse nature of AI technology: trying to find common codes would either have to be too compromised and diluted or so complex as to be incomprehensible. The other main challenge is the different cultural approaches to privacy and ethics across the globe. We can see this today between China and the US, but also between the US and Europe.
So we need to look for answers at the metaphorical coal face. I know from my work with a wide range of businesses that if the data is there then there will be value that can be drawn from AI, whether this is to improve process efficiencies or increase customer service levels, or to transform the whole business. And it’s clear to me that this value can be orders of magnitude greater than any other automation technology. But this value doesn’t have to come from the Googles or Microsofts, or be implemented by the Big Four consultancies. AI, in my mind, will work best if it is a grass roots movement: it will find its champions, advocates and sponsors in everyday business and everyday lives. This is the democratisation of AI that will unlock its value.
How can we try to democratise AI?
In the world that I am describing, AI will succeed as a business tool only where it can be used easily and effectively. For the UK to become a go-to source of AI capability, and a central hub of trade in AI, then we really need to focus on the people, both the users and the creators of the technology.
We need business people who understand the capabilities and value of AI, and we need data scientists and developers that can build it. In last year’s report on skills from the All Party Parliamentary Group on AI there were good recommendations around how to do this and these should certainly be pursued. But fundamentally we need to make AI less sexy and less scary for people. Once AI can be seen as a tool – once it has been democratised – then that is when we will see the real value flow.
AI for all
How do we make sure these AI tools are available to as many people as possible?
Open source is one answer, and here the big tech firms are providing many of their development tools free of charge, but of course as a hook to become the de facto standard. We need to think carefully about the relative roles of ‘Big AI’ versus ‘Small AI’, as there is so much energy and talent outside of these big firms that should be encouraged and promoted. A democratised AI economy will rely much more on self regulation and domestic ethical codes. We need to give the small firms the voice and the breathing space to flourish in this environment.
This leaves the chief role of government in boosting trade in AI to supporting the necessary education and skills for our people, whether this is delivered on the job or much earlier in the formal education process. This point about embedding AI into the standard curriculum cannot be emphasised enough. As everyone is aware, transforming education is not an overnight change. But, as Napoleon is reputed to have said, the best time to plant a tree is 25 years ago. The second best time is now.
To read more about the advance of Artificial Intelligence, sign up for our free newsletter.
*Photo: Chris Gayner.