• November 30, 2022

New Galaxy S23 Ultra Leak Reveals Samsung’s Significant Upgrades

Samsung’s forthcoming Galaxy S23 Ultra will gain some significant improvements over the current Galaxy S22 Ultra, according to the latest leaked info. Taking to Twitter, persistent leaker Ice Universe has revealed …

Organizations That Maintain Trust In 2023 Will Thrive

The 2020s are off to a tumultuous start. Individuals have experienced extraordinary political and social upheaval, war on the European continent, the reemergence of infectious diseases, financial instability, scandals, and extreme …

Former Spectrum Strikers Sue Union

Nearly six years ago, around 1800 cable technicians based in the New York City area went on strike against Charter-Spectrum, who they accused of union-busting in the years following Charter Communications’ …

Artificial intelligence just doesn’t pop up when you install tools and software. It takes planning and, most of all, it takes data. But getting the right data to make AI and machine learning algorithms — and understanding it — is where many organizations are slipping up, a recent study finds.

Organizations face difficulties with data silos, explainability, and transparency, a study of 150 data executives commissioned by Capital One and Forrester Consulting finds. They say internal, cross-organizational, and external data silos slowed machine learning deployments and outcomes. A majority, 57% of respondents, believe silos between data scientists and practitioners inhibit deployments, and 38% agree that they need to break down data silos across the organization and partners. More than a third, 36%, say working with large, diverse, messy data sets is a challenge.

Data may well be the Achilles Heel of AI, industry observers agree. There’s a dearth of data literacy that is slowing the pace of progress, says Ajay Mohan, principal and AI and analytics lead at Capgemini Americas. Such literacy, he explains, is “an understanding of the value of data and how to manipulate and use it to generate value.” The issue for many companies, he points out, is they “often lack the appropriate resources, such as data scientists, data engineers or technology-oriented subject matter experts to look at the business challenges and the potential for data to unlock solutions to these challenges.”

In addition, it’s often difficult to engage in another data-driven activity, articulating business value or ROI. “This also a core competency that many end-users lack,” Mohan says. Add to the mix “challenges leveraging data from disparate legacy systems and sources which can make it cost-prohibitive to develop truly meaningful AI applications.”

With a lack of data literacy come the data silos that also inhibits AI. “Even if companies had the resources to be data literate, a big challenge many larger businesses face is that of operational silos — business functions, geographic teams or other business lines operating in isolation from their intra-company peers,” Mohan says. “For example, many large consumer goods companies may operate dozens or hundreds of brands globally, each with their dedicated marketing teams and each leveraging their own technology framework and processes.”

Advertisement

There is also “inertia within companies” to traverse in the direction of data literacy, connectivity, and human skills. “They are either investing and not seeing the value or not investing enough time and money in data management systems to make it successful,” says Krishna Tammana, chief technology officer at Gupshup. “The pre-requisite to efficient AI is high-quality data- — an area many companies lack in.”

At the same time, opening up data full force to AI systems may be troublesome, introducing bias and misguided information. “We view every adoption of new AI breakthroughs through the lens of ethical responsibility,” says Peter Gordon, global head of AI product for Hogarth Worldwide. “We look very closely at how to safeguard against misuse and the harm that can be caused from algorithms that have and inherent bias in the training data. This is more of a due diligence than an issue, but rightfully will hold back more rapid adoption.”

Not everyone is missing out, and there are some use cases in which data has been successfully leveraged. “There have been some targeted use cases with right quality of data that’s working well,” Tammana relates. “One of the sectors where we see it gaining a lot of traction is customer engagement through conversational AI. The current landscape for customer engagement leaves a lot to be desired. With AI powered personalized conversations, we have seen better conversions, retention and brand recall.”

Getting a full grasp of the data needed to assure greater accuracy in output will open doors to more advanced forms of AI. “The next transformation is generative AI – the use of data to automatically generate new images, videos, headlines, music and even third worlds that have never been seen or heard before,” says Gordon. “It’s incredible to experience, and the pace of maturity is exponential. And we approach this enthusiastically, but with caution, to ensure it is ready to deploy at scale.”

Advertisement

Leave a Reply

Your email address will not be published.