IBM: Generative AI Can Make Better Decisions than the Human Brain

At the beginning of the year, IT giant IBM announced the launch of its new partner program, dubbing it IBM Partner Plus. IBM launched the program to extend partner access to various sales, training, marketing resources and incentives.

All this, IBM says, was “tailored support to deepen partners’ technical expertise and help speed time to market.” The program, IBM also noted at the time, was central to the company’s hybrid cloud and artificial intelligence strategy, which consists of tools such as Watsonx. Launched in May, Watsonx is an AI and data platform with an AI assistant designed for scaling AI across the enterprise.

IBM’s Steve Smith

“Though an emerging technology, partners have been receptive to generative AI,” Steve Smith, general manager of IBM’s global systems integrators, told Channel Futures, noting that IBM provides enablement on how to use AI and how to build the best solutions based on the tool.

“We want to move from 40% partner sales to 80%,” Smith said.

Doing so might be closer to reality than ever before, as in September, IBM Announced the general availability of ‘Watsonx’ Granite Model Series, a fresh iteration of Watson. It is a collection of generative AI models that help users infuse more generative AI into business applications and workflows that it hopes partners will sell and advocate for in greater quantities.

Generative AI and Decision-Making, ‘Better than Human Brain’ Expert Says

January of this year saw the realization of IBM’s enhanced partner portal, with the tech firm adding that it was doing all this because it was “more serious than ever about putting partners first.” The program sought to help partners widen their market opportunity and create unique revenue streams, enabling them to resell IBM products via other cloud marketplaces.

With tens of thousands of partners enrolled in its program, IBM’s Partner Plus program has been live for nearly a year. As such, we asked IBM partners to understand where AI fits in the grander scheme of their overall strategy and how they are adjusting to the bleeding-edge technology.

Wipro’s Keri Dawson

“Partners can benefit from generative AI, especially when it comes to areas like marketing, decision-making, etc.,” noted Keri Dawson, head of solutions and offerings at IBM partner Wipro and a former AWS executive. “Increasing innovation and business functions in domains that benefit from creativity, like marketing, and improving decision-making, large language models (LLMs) can execute analytic decision-making better than the human brain.”

Security Guardrails, Bias and AI Hallucinations

Noel Hara, vice president and chief technology officer, public sector, at NTT Data Services, an IBM partner, closely watches AI for his company.

Security is always of concern, namely when dealing with AI. So it makes sense that the quality of data fed to generative AI models matters, Hara told Channel Futures. In the public sector, it must be government-approved and packaged into a container.

NTT Data’s Noel Hara

“Have you ever been on page five or six of a Google/Bing search? There are some scary things out there,” Hara said.

Simply put, said data should be limited and adhere to FedRAMP guidelines. It is not publicly available data – from which ChatGPT pulls – but rather refined data that likely doesn’t lead to a lot of bias and/or hallucinations as the appropriate security guardrails are in place.

“Bias has to be extracted beforehand, and humans have to remain in the process,” said Hara.

Wipro’s Dawson then gave an example of a tech firm that saw the potential pitfalls of generative AI firsthand, something commonly presented in a favorable light.

“There will be bias, and to pretend there will not be is just not practical, as generative AI constantly generates new data,” Dawson said, telling the story of how generative AI took on a mind of its own.

Management at a small tech firm laid off some of its development staff to have generative AI pick up the slack, looking to save time, resources and money, only to discover the AI had built a backdoor to the code “just in case.”

“We don’t know what the AI will do as it is always spawning novel data/content, which is where ‘hallucinations‘ come from,” Dawson said.