(As originally published on LinkedIn, 27 March 2018)

Why AI’s biggest skill gap isn’t technical

  • Share

By: Cathy Cobey, Partner, Technology Risk at EY Canada

To respond to the chronic shortage of artificial intelligence (AI) talent, universities and technical schools are responding with new and expanded programs in data science and AI. However, the biggest skill gap in AI isn’t with technicians – it spans across the organization.

That may be a surprising statement to some, as the gap in AI technical skills is large and growing at an alarming rate. In its 2017 U.S. Emerging Jobs Report, LinkedIn reported that machine learning engineers and data scientists had risen to the top two emerging jobs. And in a recent global EY poll of senior AI professionals, 56% of respondents indicated that a lack of AI experts is the greatest barrier to AI implementation within business operations.

That said, businesses are continuing to integrate AI into their business operations despite a lack of talent. The EY poll found that 67% of respondents have someone specifically dedicated to driving AI efforts inside their organization. With sentiment shifting to a greater acceptance of AI, businesses are overcoming internal resistance to the technology.

This shift in business acceptance is creating the biggest skill gap for AI – a lack of AI-knowledgeable business leaders that can develop an effective strategy, implementation plan and governance structure for AI. As noted by Nigel Duffy, EY Global Innovation AI Leader, “To be successful, leaders should identify a business challenge and then determine where the technology can solve this problem – but this can only be realized with the support of AI-savvy professionals who can identify AI opportunities in their business.”

In our fourth video of our Managing the Risks of AI series, Cindy Gordon, CEO of SalesChoice, and I discuss the risks of organizations forging ahead with AI without adequate governance and outline several ways in which organizations can become AI ready.

Starting at the top, there is a need for the C-suite and Board of Directors to educate themselves on the opportunities and challenges of AI beyond the boardroom. They need to immerse themselves into the design experience by visiting their innovation labs to understand what is being developed, the development methods being used and who are developing them.

They also need to work with their risk and control functions to develop a set of design principles, including operational boundaries, that the organization is comfortable operating AI within. These could include:

  • Restrictions on the level of autonomy that the AI can operate – augmented or full
  • Guidelines on the use of AI methods for regulated, financial or customer interactions
  • Standards for minimum confidence thresholds and/or algorithmic transparency levels
  • Protocols for data acquisition, cleansing, use and storage
  • Ethical or moral norms to be followed

The challenge in developing the above operational boundaries is that new methods and use cases for AI are emerging every day. Effective stewardship over AI will require a dynamic governance model. Although some organizations have responded by creating an Innovation Advisory Council or appointing a Chief Ethics Officer, to be truly effective the application of ethical design principles will require day-to-day application by those sitting at the AI design table.

Both technicians and business representatives need to be equally versed in the organization’s design principles, operational boundaries and ethical standards. To address the bigger AI skill gap across the organization, businesses investing in AI need to develop enterprise-wide training programs. In addition, academic institutions need to demonstrate the same level of urgency in incorporating AI into their mainstream programs and the development of technical AI programs; while professional organizations need to incorporate intelligent automation into their foundational or core knowledge requirements.

Also, it goes without saying that the design principles and operational boundaries also need to be codified into organizational polices, standards and procedures and supported by an effective accountability structure.

Watch our fourth video on Managing the Risks of AI for a more in-depth discussion on the need to address the AI skill gap at both the technical and enterprise-wide level.