CIOs interested in moving forward with the technology are now working hard to clean up and manage their data so they can take full advantage.
Bala Krishnapillai, vice president and head of the information technology group at Hitachi Americas, said the organization has encountered instances of inconsistent, duplicated, and incorrect data, leading to contradictory information that confuses AI outputs.
He said the company is regularly updating and refining its data to ensure accurate results from AI tools accessing it. That process includes the organization’s data engineers validating and cleaning up incoming data, and curating it into a “golden record”, with no contradictory or duplicate information.
Google Cloud Chief Evangelist Richard Seroter said he believes the desire to use tools like Gemini for Google Workspace is pushing organizations to do the type of data management work they might have been sluggish about in the past.
Isabelle Bousquette
If you don’t have your data house in order, AI is going to be less valuable than it would be if it was, he said.You can’t just buy six units of AI and then magically change your business.
An entirely foreseeable obstacle for anyone working in operations, as opposed to top management who get their information about the company’s internal mechanisms neatly delivered to them in a glossy presentation. I have been through enough jobs to know that most large companies are built on a jumble of disconnected legacy systems that rarely communicate seamlessly. Also, most C-level executives are reluctant to invest into upgrading and modernizing the data infrastructure, fearing disruptions to regular operations and sunk costs without tangible benefits. Expecting an AI system to leapfrog these issues and deliver perfect results with no human input was always a rose-tinted proposition – if not downright delusional.
Especially if you think about the multiple times AI simply hallucinates information that sounds plausible instead of admitting it doesn’t have an answer. The current general purpose chatbots have not been trained to handle financial data and compliance, or to make plausibility checks on their results, something any remotely competent person would do before submitting critical reports. I didn’t expect this hype around AI in the workplace to deliver concrete changes anytime soon, and it’s refreshing to see companies starting to acknowledge that.
The core issue for the tech companies promoting LLMs as productivity tools though is that their main selling point is precisely AI as silver bullet: buy our off-the shelf assistant and get improved productivity overnight – with a sprinkle of layoffs on the side. If companies must go through extensive data management and process reengineering to see benefits, they might hold off on the LLM contract until after putting their systems in order. They might even decide that the internal streamlining delivered enough gains to forego expensive subscriptions for unproven tools altogether – a disaster scenario for AI companies, which after investing staggering amounts into chips and training models would find themselves without viable products on the market to generate revenue.
Post a Comment