Buoyed by predictions from firms such as PwC that estimate artificial intelligence (AI) could add $15.7 trillion to the global economy by 2030, enterprises are well aware of the potential of AI technologies to improve efficiency, accuracy and reduce overall expenditure.
Yet many are faced with significant challenges in adopting the technology. AI Business recently surveyed 500 Fortune 100 C-suite business leaders with the aim of deciphering what is preventing AI from really taking off in business.
The very first obstacle identified by the vast majority of respondents may come as a surprise to those irritated by the substantial media coverage surrounding AI. More than 92 per cent of respondents identified the lack of understanding about AI, its capabilities and overall potential as the number-one issue preventing them from making the most of the technology.
A lack of understanding central to the sluggish uptake of AI
“The demystification of AI is vital, otherwise fear about what the technology can do to jobs, skills and people will proliferate,” says Sandeep Dadlani, Mars Inc’s chief digital officer. “There is also this expectation that AI will produce magical results in a short period of time. That reality is far away – the truth is that AI is still in its infancy in terms of solving problems.”
Most organisations not only lack the necessary knowledge, but also the right talent to make a success of AI; an issue that nearly all members of the C-suite cited as a major hurdle. The race to recruit top AI talent has reached unprecedented levels among the industry’s leading tech vendors, to the point where even renowned universities are struggling to retain their staff on oversubscribed machine-learning courses. As a result, it’s even harder for enterprises to compete for AI talent, which is why many organisations have taken the route of outsourcing AI to external vendors.
AI ownership pivotal to successful implementation
Closely tied to the talent deficit is the pivotal issue of ownership of AI within the organisation, identified by nearly three quarters of all respondents. There’s no clear path to AI implementation, with some companies choosing to run small, case-by-case projects and others opting for a comprehensive AI or digitalisation strategy.
Research consultancy Tractica recommends that centralised AI control is best for organisations on an advanced AI path, with multiple implementations at scale, while most businesses starting out in AI are better off with decentralised projects.
For many large enterprises, it falls to digital or innovation-oriented members of the C-suite, such as chief information officers (CIOs), to evangelise for the technology within the wider organisation, but building a coalition of the willing, based on clear use-cases, is vital.
“It can’t just be the CIO walking into the executive meeting with AI,” argues Sherif Mityas, chief experience officer at TGI Fridays. “You need to work with those who have been through the business case. There can be lockstep between the CIO and other board members who are able to see the impacts of AI in their own departments and areas of responsibility.”
Data is the fuel and the starting point of AI
The shortage in AI talent and literacy, as well as a nebulous implementation structure, leads to another significant obstacle of in-house versus outsourced AI. With understanding of machine-learning murky at best among executives, selecting the right AI solution remains challenging for businesses. Cloud service providers such as Amazon Web Services and Microsoft Azure offer enterprises a base with which to clean their data and develop their own AI solutions, while a constantly growing base of AI startups offers specialised, use-case-oriented applications.
There’s no one-size-fits-all approach, which is why many AI vendors recommend starting with a specific business problem, rather than looking to apply AI for its own sake. “Business sectors should assess and align critical business problems with opportunities for AI technology to provide value through cost avoidance and/or new capability development,” says Andrea de Souza, global business development lead for NVIDIA. Ms de Souza recommends that businesses begin with one, two or three practical experiments to explore if AI could have a significant impact for their organisation. The key is to combine in-house and outsourced approaches to implementation according to circumstances and need.
The most persistent technical issue that prevents rapid deployment of AI in the enterprise is undoubtedly organisational data. Data is the fuel of AI and, in many cases, its starting point. Without a good data strategy and well-organised datasets, businesses will be unable to implement AI efficiently or effectively. Thankfully, enterprises are growing savvy to this, with many of them now investing heavily in their own data scientists and experts. For those operating on a lean business model, there exist plenty of third-party consultancy services to help with the heavy data lifting.
The UK can become a standard bearer for ethical AI
The elephant in the room is undoubtedly public trust. More than 80 per cent of our respondents remain concerned about customer and societal responses towards putting AI to work. More than 60 per cent also highlight worries about employee attitudes to working alongside machines. A larger conversation around ethics, AI and the overarching impact of the fourth industrial revolution has yet to take place.
Thankfully, this is one area in which the UK is an emergent leader. Parliamentary reviews into the prospects for AI in Britain have found that, while the UK might never match global competitors in spending, we have the infrastructure and know-how necessary to become a standard bearer for ethical AI. Lord Clement-Jones, chair of the Select Committee on Artificial Intelligence, says the UK has many of the right ingredients in place. “Businesses should focus on ensuring government delivers the climate for AI in terms of getting the context right for growth,” he says. “We don’t think a special AI regulator is the way forward. It’s really important that the tech industry gets behind all those principles around transparency and explainability, because we mustn’t lose public trust.”
The road to AI is long and bumpy, and uncharted territory. However, the opportunity is too significant to ignore. Businesses that don’t adopt AI will be left behind in the long run. AI is already enhancing human productivity, yet the long journey to an AI-powered economy has only just begun.
