How AI is democratising business data

Experts predict that generative artificial intelligence (AI) will make data available to employees of all kinds. With the democratisation of data comes serious governance and security risks.

In the MIT Technology Review’s Laying the Foundation for Data- and AI-led Growth, a range of tech, data, and information experts shared their thoughts on what business leaders should know about the new technology in 2024. The report unearthed a range of findings, including:

• Chief information officers are “doubling down” as every organisation they surveyed plans to boost its spending on AI and data this year, while nearly half will increase by more than 25 per cent.
• Executives expect big things from AI, as 81 per cent of survey respondents expect AI to boost efficiency in their industry by at least 25 per cent in two years, while one-third predict boosts of 50 per cent or more.
• The largest share of respondents (39 per cent) said investing in talent is what needs most improving when it comes to data strategy.

In this article, we’re focusing on a different finding: the power of AI democratisation. “Six months ago, I would have said democratisation of AI is probably a decade away. But with generative AI, it’s happening in front of us,” said Jon Francis, chief data and analytics officer at General Motors.

Cloud democracy

“Generative AI is the first big step in AI democratisation,” said Naveen Zutshi, chief information officer at Databricks. Understanding what this means requires a bit of history. As noted in the report, the term “democratisation” first featured in the enterprise technology conversation when cloud technology first surfaced.

Cloud computing services, like Amazon Web Services, Microsoft Azure and Google Cloud Platform, allow tech and software businesses to access pre-built tools to save them from having to build a back end from scratch.

“They no longer have to spend all their time building a back end because those components are off-the-shelf features in the cloud,” explained Forbes.

“This shift frees up teams to focus on the business problem they want to solve and the overall user experience rather than the technical challenges along the way.”

What this means is smaller companies can build on existing software without the kind of upfront investment that had prevented them from edging in on software giants. The same phenomenon was observed within businesses as employees in different organisational units were given easy access to work applications.

AI and data democracy

The same change is prophesied for data as the business value of allowing more employees to access data and insights derived from analytics became clear. Sixty-four per cent of respondents to the MIT Technology Review survey consider building the data architecture needed to enable secure sharing of live data across platforms to be “very important” in achieving their technology aims.

Similarly, 41 per cent of respondents consider it “very important” to have a managed central marketplace for datasets, machine learning models, and notebooks. But enough of the jargon.

We’ve all heard of ChatGPT and similar text and voice generative AI models. Jeffrey Reid, chief data officer at Regeneron Genetics Centre, anticipates that these models will replace the specialised services used by software developers today: “A more casual interface will make it easier for everyone, even those with no coding skills, to query data.” But the benefits, he said, go beyond analytics.

“It provides an opportunity to help you remember something, or to validate something that you already know, in order to support decision making. We’re seeing that enable everyone across the board,” he said.

Previously, Gartner anticipated that “every business will be an AI business”. With the broader democratisation of generative AI, it’s plausible that every employee, too, will become an AI user.

The risks

With broader access to and use of data and generative AI tools come new risks. Though no one is entirely prepared when it comes to mitigating AI risks, it’s fair to assume those with specialised knowledge are best placed. Now that the floor is opening to employees of all backgrounds, risk management is a growing threat.

“You can employ anyone at any level of the organisation today to go build an AI application,” said Mr Francis. “But are we fully thinking through all the legal, privacy, and security implications, or the commercial ones, such as what it means from a brand perspective?”

As explained by the MIT Technology Review, compliance will, in part, be guided by lawmakers and public regulation, but senior management has a role to play beyond the law. When asked what concerns them most about their data and AI preparedness, 26 per cent of executives cited “inadequate governance”. Building the required governance mechanisms will be a necessary piece of the data and generative AI democratisation puzzle.

“As business units and their staff clamour to use generative AI, executives seek assurance that governance frameworks for the technology can provide not only the needed data accuracy and integrity but also adequate data privacy and security,” said the MIT Technology Review.