Chief information officers are aware that not every tech issue can be solved using generative AI. However, others are now forced to temper the excitement of boards and chief executives who are intoxicated on the promise of the technology.
In certain circumstances, we must temper the belief that we ought to be utilizing GenAI more and more. Mr. Cooper’s nonbank mortgage servicer and originator, Sridhar Sharma, CIO, spoke of his regular discussions on artificial intelligence (AI) and generative AI with the CEO and other C-suite colleagues.
Since the release of OpenAI’s ChatGPT in November 2022, generative AI—a type of AI that creates text and other content using prompts—has captured the attention of Wall Street and corporate boardrooms, propelling AI chip manufacturer Nvidia to a trillion-dollar value.
Stock prices of companies that include mentions of generative AI in their earnings reports have risen, leading to a competition amongst businesses to be the first to deploy the technology.
Nevertheless, some CIOs said they are conscious of the urge to force the technology into uses that are better served by more traditional kinds of predictive AI or even something as basic as a spreadsheet.
Sharma stated that there isn’t a single use case where generative AI works well. In actuality, it can occasionally be hazardous and less effective. According to him, generative AI solutions can be more expensive, difficult, and prone to hallucinations than other types of AI.
Sharma stated that Mr. Cooper, for instance, has been employing nongenerative AI for the past three years to determine whether certain details, like a signature, stamp, or notary date, are absent from a given document. Additionally, the company investigated whether generative AI may have provided a better solution to the issue, but it was unable to do so.
He stated, “At this point, we see no point in trying to reinvent the wheel just because GenAI can do it.”
Head of analytics and AI at TD Bank Luke Gee said he frequently responds to questions and suggestions regarding generative AI from colleagues in the business side. He claimed that there isn’t a day that goes by that he doesn’t receive a fresh article link on something.
When he does, Gee helps colleagues understand the issues at hand and decides if generative AI is the best course of action. And, he added, it might be or it might not be.
According to Sameer Maskey, founder and CEO of enterprise AI company Fusemachines, generative AI is wonderful for content generation but less useful for tasks like demand forecasting, anomaly detection, predictive maintenance, and churn prediction.
Businesses must to examine the issue first and determine what kind of algorithm makes sense. he stated. He said that if they don’t, CIOs may be employing generative AI to appease superiors and will probably end up wasting time and money.
It can take a full-time job to train the staff, from the C-Suite to the trenches, on the right use and deployment of a technology as hyped as generative AI.
The chief technology officer of Akamai Technologies, Robert Blumofe, stated that the hype is completely taking the air out of the conversation and failing to give other kinds of AI models the proper attention that can actually produce value. The resolution of the business issue is not the aim. Adopting AI is the aim.
According to Blumofe, he’s working to inform the rest of the team about the true capabilities of generative AI, as well as when and how to apply it. According to him, Akamai, a cloud computing, content delivery, and cybersecurity company, uses more traditional nongenerative AI techniques to spot patterns in network traffic, alert users to cybersecurity issues, and suggest products to them.
According to him, these nongenerative algorithms have the advantages of being more accurate, less costly, and energy-efficient than emerging generative AI models. He claimed that their reduced stature is partly to blame for this. According to him, applying GenAI may need using megawatts to address issues that could be resolved with milliwatts.
The company’s security portfolio does make use of some generative AI techniques. The issue, he continued, is the propensity of modern times to feed copious amounts of data into massive generative AI models for tasks as basic as finding the median or the 95th percentile of a given collection of data—questions that can be readily resolved with a simple spreadsheet.
As some CIOs have stated, there must be open contact with the board.
Chief technology and platform officer at McKinsey Jacky Wright suggested CIOs advising boards to assist them in determining whether generative AI can actually boost core capabilities.
Okay, let’s get started. If not, what aspects of running your company can we concentrate on to truly help? She concluded.
According to her, during those discussions, CIOs should steer the board clear of areas where generative AI may not yet be developed enough to provide its full potential, including AI agents that finish an entire end-to-end workflow
Therefore, why concentrate on that when we may look at something else with a better value and go more quickly? She said, “This is an ongoing dialogue.”
Fidelity Wealth’s chief of technology, Eric Emerson, stated that the financial services organization has been utilizing nongenerative AI for nearly ten years. However, at this time, generative AI isn’t the solution to some of the most pressing issues the company is trying to address.
He does wish AI could help with the modernizing of our mainframes. He said, “That would be awesome.” But the answer to that is currently no.
