In association with

The global AI agenda: Promise, reality, and a future of data sharing

April 14, 2020
A new survey of 1,000 AI leaders reveals the top AI use cases, the challenges of building scale, and the potential benefits of data sharing for business and society.

Artificial intelligence technologies are no longer the preserve of the big tech and digital platform players of this world. From manufacturing to energy, health care to government, our research shows organizations from all industries and sectors are experimenting with a suite of AI technologies across numerous use cases.

Among the organizations surveyed for this report, 72% had begun deploying AI by 2018, and 87% by 2019. Yet much remains unknown about AI’s real, as opposed to potential, impact. Companies are developing use cases, but far from all are yet bearing fruit. How, business leaders ask, can we scale promising use cases to multiple parts of the enterprise? How can we leverage data, talent, and other resources to exploit AI to the fullest? And how can we do so ethically and within the bounds of regulation?

MIT Technology Review Insights surveyed 1,004 senior executives in different sectors and regions of the world to understand how organizations are using AI today and planning to do so in the future. Following are the key findings of this research:

  • AI deployment is widespread but will take time to scale. AI is being deployed widely across sectors, but its reach within enterprises is likely to expand slowly. Most survey respondents (60%) expect AI to be used in anywhere from 11% to 30% of their business processes in three years’ time, exercising an important, though not dominant, influence in their operations. Financial services providers, manufacturers, and technology companies have the highest expectations of AI penetration.
  • Change management and data challenges do most to hinder scaling of AI. Taking AI use cases beyond the pilot stage is far from straightforward for any organization. The surveyed firms struggle most with the change management involved in modifying business processes to leverage AI, a challenge cited by 51% of respondents. Nearly as difficult are data challenges—cited by 48%— chief among which are difficulties integrating unstructured data and interfacing with open-data platforms (problems reported by 57% and 53% of executives, respectively). Respondents’ emphasis on the latter suggests a desire to access external data to feed their AI models.
  • The top AI use cases today are in the areas of quality control, customer care, and cybersecurity. Around six out of 10 manufacturers and pharma companies are using AI to improve product quality. Nearly half of consumer goods and retail firms (47%) are using it in customer care. Over half (51%) of energy firms are leveraging AI for monitoring and diagnostics, 58% of financial services providers for fraud detection, and 52% of tech firms to strengthen cybersecurity.
  • Currently nascent, data sharing can magnify the impact of AI. Two-thirds (66%) of surveyed firms are willing to share internal data externally to help develop new AI-enabled efficiencies, products, or even value chains. Manufacturers, consumer goods firms, retailers, and health-sector organizations envision benefits to supply chain speed and visibility and reduced time to market for new products. Technology and financial services firms see gains to customer service, cybersecurity, and fraud detection, among other uses. Businesses are still cautious, however, and more clarity is needed in privacy regulation and industry standards, say 64% and 58% of respondents, respectively, before data sharing takes root.
  • Early AI adopters are benefiting the most, but also have war stories. Firms with the longest experience of using AI are learning by trial and error and ultimately benefiting. Surveyed organizations that first deployed in 2015 are more likely than those deploying later to have seen their AI projects underperform in terms of return on investment (ROI). But the early adopters are also more likely than others to say that ROI has exceeded expectations.