Artificial intelligence skills shortages re-emerge from hiatus

Two in five companies see lack of technical expertise as a roadblock to AI. It couldn't come at a worse time.
Written by Joe McKendrick, Contributing Writer

Back in June, a report from LinkedIn noted that the Covid crisis had cooled off demand for artificial intelligence skills. However, four months later, companies are still struggling with finding AI skills. 

Photo: Joe McKendrick

Overall, more than four out of 10 enterprises now use artificial intelligence in a serious way, up one-third in just two years, a recent survey of 1,000 executives by RELX finds. Adoption accelerated in a big way over the past few months. AI technologies are being employed at 81% of businesses -- up from 48% in a similar survey conducted in 2018. 

At the same time, AI talent is in short supply. The leading reasons for companies not using AI are budget constraints (44%) and lack of technical expertise (39%).

Advanced AI adopters are seeing the most acute shortages of talent, cited by 23% in a recent Deloitte survey. It is possible that the more a company works with AI, the more likely they are to understand what skills they actually need, says David Jarvis, a researcher with Deloitte. "It could also be that the 'seasoned' tend to pursue more transformational projects using AI, focusing more on creating new products and services than on cost reduction," he adds. 

Companies were under the gun to quickly ramp up AI efforts as they shifted into crisis mode over the last seven months, the RELX survey shows. The majority of respondents (68%) increased their investment in AI technologies during the Covid crisis, with 48% investing in new AI technologies and 46% investing further in AI technologies already in use at their companies.  Similarly, 63% of business leaders polled report that AI technologies had a positive impact on their business's ability to stay resilient during this time. 

There also appear to be efforts to bring professionals up to speed in what is needed to build and deploy AI-driven systems. The RELX survey shows 75% of companies now offer training in AI technologies -- up from 46% in 2016. 

As Deloitte's Jarvis points out, the type of talent most in demand-AI developers and engineers, AI researchers, and data scientists. Most don't turn to training, but prefer to look for experienced AI professionals from outside their organizations. Still, the more seasoned AI adopters have learned that internal development best answers the need for skills -- "seasoned adopters rely more on internal employees already trained in AI; Starters lean more on partnerships with other companies that have AI expertise." 

Roughly two-thirds of this group is currently training their developers to create new AI solutions (64% compared with 43% of starters) and training their AI staff to deploy AI solutions. They are also providing training for employees to use AI in their jobs (67% compared with 48% of starters). 

"Diversify sources of AI talent," Jarvis advises. "Look at how experienced hires, university hires, and partners and vendors can help fill gaps. Aim to build a bench of business talent that can speak AI as well."

Culling through AI job descriptions in recent classified ads provides evidence of the skills currently in demand: 

  • Senior data scientist (financial services company): "Develop the next generation AI/ML-enabled financial advice solutions. Apply analytics, artificial intelligence, machine learning, and cloud technologies knowledge to create foundations for the next generation of smart and cloud connected platform and services. This work will transform how financial advice is made accessible to the masses."
  • AI/ML engineer (pharmaceutical): "Machine learning experts that want to be part of a team that discovers new medicines. Influence machine learning strategy for a program/project; Explore design options to assess efficiency and impact, develop approaches to improve robustness and rigor  Create algorithms to extract information from large, multiparametric data sets; deploy algorithms to production to identify actionable insights from large databases." 
  • AI software engineer (technology company): "Help develop an acquisition system that takes full advantage of AI. Enhance and streamline the end-to-end process by utilizing new and emerging technologies such As Blockchain, AI and Robotic Process Automation. Must have UI and Angular JS experience." 
  • AI/ML product designer (consulting company): "Need to have experience building functional applications and developing UI specifically for AI or ML tools. Work with team to understand feature functions of NLP tools, execute rapid prototypes, and deploy. This role will be interacting with the data. Help with UI design; bridge the gap between design and end user; come up with the design for the end users interaction with AI tools."
  • AI engineer (FinTech):  "Role bridges the gap between software engineer and research scientist. Develop solutions for real-world, large-scale problems using AI/ML; improve deep learning engine that services clients. Add more skills to the neural network, define training and validation sets; code automated tests; setup feedback learning loop and fine-tune behavior based on customer interactions; formulate and run experiments to improve comprehension rate by altering topology, tweaking activation and learning function and alike; assist in coding execution engine." 
  • AI machine learning engineer (aerospace manufacturer): "Possess passion for conceptualizing, building, testing and maturing autonomous systems. Apply technical standards, principles, theories, concepts and techniques in the software field to develop tangible, differentiated and meaningful solutions. Focus on development, integration and test, spanning the complete system development lifecycle. Includes requirements generation, system and software design, implementation, integration, and test."
  • Program manager -- AI/ML strategy and governance (Financial services): "Increase the volume and velocity of AI applications. Lead programs to establish firm-wide level AI/ML processes and governance from development to implementation. Work across organizational boundaries, articulate views/roadmaps, and understand the interaction between various initiatives to deliver solution. Influence senior stakeholders to define and agree on a program delivery framework."
Editorial standards