MIT’s Sloan CIO Symposium always provides a strategic, high level overview of the business and technical challenges facing chief information officers. At this year’s event the central role of data, the specter of increased rules and regulations and of course everything AI were top of mind.
Among the highlights:
Governance and management of fast-changing technology. Look for more rules from more sources governing IT. For example, MIT Professor Stuart Madnick pointed out that California is the first state in the country to implement a law mandating security for devices on the Internet of Things (IoT.) While security experts have long urged ransomware victims not to pay the hackers, lest they encourage further attacks, North Carolina is the first state to prohibit its agencies and local governments from making such payments, with others possibly on the way.
If it isn’t the government telling you what you can’t do, it might be the insurer from whom you buy cybersecurity insurance. Such insurers will often insist a policyholder leave the hacked systems undisturbed and offline until their security experts determine what happened, said Jeff Reichard, vice president of solution strategy/product strategy at backup and recovery software vendor Veeam. That means companies need a separate backup infrastructure to keep the business going in case their insurer turns their primary system into a virtual crime scene.
Whether and how to control the fast-growing abilities of generative AI was another hot topic. Alex Singla, senior partner at McKinsey & Co., predicted the accepted wisdom that AI is a “black box” whose workings are impossible to explain “won’t fly” within seven to eight years. That means users must make AI a “glass box” where shareholders, customers and governments can understand how AI made its decisions.
Beyond the AI Hype
Amid all the talk about the possibly wonderful or horrific uses of generative AI, several attendees described practical benefits they’re seeing right now. Thomas Peck, executive vice president and chief information and digital officer at Sysco, says the food distributor is using AI for, among other things, suggesting products to cross-sell and up-sell to its customers and to predict shrinkage (theft) and inventory levels. Another attendee said his firm is using generative AI to find the best pitches, voice mail and email messages for prospects, as well as the most qualified leads to leave these messages with.
Brightspeed Chief Technical Officer Ashok Kumar described how the Internet provider is using AI to use information as specific as the location of utility poles to “design an entire network in a matter of minutes” compared to the weeks that work typically requires. Bottom line: Rather than wait for global insights about what AI means for the future, grab all the benefit you can as soon as you can.
It’s Your Data; Take Care of It
AI is the ultimate example of GIGO (garbage in, garbage out), to use a time-honored phrase circulating around the symposium halls. AI trained on inaccurate or biased data can quickly produce industrial-scale havoc.
Just as some organizations insist on a bill of materials (BOM) for the software they use to better find potential malware, they may want to create similar BOMs for the data used to train generative AI, said Madnick. This not only helps ensure its accuracy but identifies who generated the data and is liable if it results in harm from poorly trained AI models.
The right data, if properly stored and curated, can uncover new markets. Girish Hoogar, global head of engineering, cloud and software at Lenovo, described how the hardware maker created its Ideapad notebook brand when customer data revealed how many customers were using inexpensive notebook computers for gaming.
That’s why data curation – ensuring it is accurate, up to date, properly cleansed and stored in a form it can be easily understood and accessed– is so important.” If we didn’t have the proper cataloging or tagging of (customer) data, we wouldn’t have ended up with the Ideapad,” he said.
If data is a digital asset, it must be maintained like a physical asset – a job that never stops. “So many times, data is treated like a public works – nobody wants to pay for it,” says John Spens, director of data and AI North America for technology consultancy Thoughtworks. He and others at the symposium described various decentralized, mesh, or federated data architectures that ensure those who rely on data the most also do their part to pay for its ongoing care.
Career No Longer Over?
At one point, the joke was that CIO meant “career is over” when big technology investments failed to deliver returns or critical infrastructure crashed. But as technology becomes an ever greater driver of revenue and profits, and creates entire new markets, one speaker predicted CIOs can become “the architect of the future of the company.”
But before doing so, they must keep multiple plates spinning. These range from keeping the day to day business runing smoothly, overcoming internal resistance to data sharing, and being the first to use powerful new technologies such as AI for good, not evil.
For drill downs on these topics, check out more coverage of the symposium here.