Technology expert Evan Schuman takes an authoritative look at the faults and foibles of enterprise IT.
The criticisms aimed at the technology — the lack of reliability, data leakage, inconsistency — offer a playbook for growing a business that competes with rival companies that leverage AI.
When Anthropic cancelled the AI account of a Swiss company that depended on the service, the move was entirely automated. A lawyer got involved and the account was restored within a day — minus 80% of the data. Oops.
‘You wouldn't let a hallucinating employee make 10,000 consequential decisions per hour with no supervision,’ says one consulting firm exec. ‘Stop letting your AI systems do exactly that.’
The key reason: most enterprises rely on pretty much the same disaster recovery plan they’ve used for years — even though their environment has changed dramatically, thanks to SaaS, cloud, and AI.
Big AI players are insisting that enterprises share with them every bit of sensitive data. But they have yet to show they’re especially good at handling the data they already have.
The company listed all of the systems that went haywire, but never really identified what happened differently that day to cause problems. Worse yet, putting out technology brushfires this way leaves the forest in danger of burning again.
Deloitte is the latest firm to discover that blindly trusting generative AI is a suicide mission.
Most layoffs related to generative AI are either the result of overhiring efforts or overzealous cuts that will be reversed in three months. But Oracle might have figured a real rationale for cuts
IT leaders need to protect AI investments from a potential bubble burst — without knowing when it will come or what it will look like.
Developers have every reason to not document their work unless they are forced to — and IT managers have every reason to not make them do it.
Over the next couple of years, automated bot shoppers are likely to start replacing human shoppers. And that’s going to force retailers to shift sales tactics to be successful.
Cloudflare’s latest maneuver is to give enterprises a choice: they can block genAI bots or charge them. But if you make a money grab now, you might soon regret it.
The Web cert world has undergone a lot of change this year, and Google has now set a June 2026 deadline for being even stricter about which certificates can be used for which functions.
Given that genAI hallucinates and ignores guardrails, do we really want it to have access and control apps from PayPal, Stripe, Shopify, Square, Slack, QuickBooks, Salesforce and GoogleDrive?
Despite how expensive genAI is right now, there’s an assumption pricing will soar in a year or two — as soon as enterprises are so invested in the tech that almost any price hike will cost less than making a switch.
Three recent incidents are reminders that generative AI tools remain troublesome and unreliable. IT buyer beware.
Online generative AI robots that scrape content and ignore websites’ ‘Do Not Enter ‘signs can cost enterprises a lot of money. But the underlying issue is how those bandwidth charges are calculated.
The Mayo Clinic has been experimenting with forcing the technology to reveal source links for everything. And there are other approaches that might be viable.
As the reliability of generative AI data remains doubtful, IT leaders need to seriously consider their risk tolerance. Most corporate boards certainly won’t.
Whether it’s genAI, new cybersecurity concerns, or the actual threats from Quantum computing, there’s a lot of misinformation bouncing between the board, C-level execs and various business units. IT leaders now need to be tech explainers.
Sponsored Links