Technology expert Evan Schuman takes an authoritative look at the faults and foibles of enterprise IT.
Non-English generative AI models are far less accurate and useful than their English-based counterparts — sometimes dramatically so. Companies paying for them should know what they’re getting, and push for price cuts.
If IT execs want to avoid doing business with genAI firms that help make weapons, that demand needs to go into RFPs. It’s the only effective way to change policy.
The glitches that shut down American Airlines on Christmas Eve and stopped all train traffic in Norway on Christmas Day both involved third-party IT vendors. But how the two enterprises handled each mess was very different.
When Macy’s on Wednesday reported more details about the “hiding” of $151 million, it became clear their accounting controls simply didn’t work. It exposed a massive software hole in just about every enterprise environment.
The TikTok owner fired — and then sued — an intern for ‘deliberately sabotaging’ its LLM. This sounds more like a management failure, and a lesson for IT that LLM guardrails are a joke.
Enterprise CIOs today allow any user to choose pretty much any freeware browser they want — and then use it to access their most sensitive systems. Does anyone see a problem here?
In a perfect universe, the persuasiveness of an argument would not be based mostly on who said it. In the world we live in, though, it is. And it’s hard to find a less credible entity to create a genAI accuracy test than OpenAI.
A group of Harvard students experimented with AI-linked eyeglasses, offering a powerful peek into the AI nightmares coming for IT in 2025.
The adoption of generative AI is moving too quickly — and its dangers remain too unknown — for any meaningful rules to be put in place on AI vendors. Regulating enterprises makes far more sense; influence enterprise behavior and the vendo....
Just about every generative AI vendor offers enterprise CIOs all kinds of promises about the technology. But they’re talking up 2026 capabilities when trying to make 2024 sales. That’s a recipe for disaster for both buyer and seller.
It’s all a matter of understanding how your business can benefit from generative AI tools and platforms. But first, you need to make some difficult decisions — and then hope genAI doesn’t self-destruct.
CIOs are so desperate to stop generative AI hallucinations they’ll believe anything. Unfortunately, Agentic RAG isn’t new and its abilities are exaggerated.
In many ways, the rush to try out still-evolving generative AI tools really does feel like the Wild West. Business execs need to slow things down.
Generative AI advocates say genAI tools can catch errors made by other genAI tools — but humans must still check the AI checkers’ work.
If you can't trust the product, can you trust the vendor behind it?
Corporate privacy policies are supposed to reassure customers that their data is safe. So why are companies listing every possible way they can use that data?
It’s bad enough when an employee goes rogue and does an end-run around IT; but when a vendor does something similar, the problems could be broadly worse.
Given the plethora of privacy rules already in place in Europe, how are companies with shiny, new, not-understood genAI tools supposed to comply? (Hint: they can’t.)
Why are so many companies sending out emails to customers that look like phasing attempts? Don't they pay attention to their own security efforts?
When McDonald's in March suffered a global outage preventing it from accepting payments, it issued a lengthy statement about the incident that was vague, misleading and yet still allowed many of the technical details to be figured out.
Wall Street’s obsession with quarterly earnings has made it extraordinarily difficult for most enterprises to spend on long-term investments, or even mid-term investments.
Ever use one of those mobile food delivery apps — only to realize your delivery person isn't who you expected? There's a lesson here about identity, authentication, and what happens when the best laid tech plan meets human beings.
The IT community is freaking out about AI data poisoning. For some, it’s a sneaky backdoor into enterprise systems as it surreptitiously infects the data LLM systems train on — which then get sucked into enterprise systems.
The New York Attorney General’s office sued Citibank for failing to reimburse customers victimized by fraud, raising serious issues all enterprises must figure out. When should a customer be reimbursed for fraud? And at what point do a customer....
One of the oldest and most frustrating rules about email spam is that the unsubscribe link never works — all it does is confirm your email address is active. But what if the unsubscribe failure is caused by something far more problematic?
It's no secret that enterprise IT in recent years has been disappointed in corporate clouds. But in general they've not done anything about it. That could soon change.
As generative AI fever continues to mesmerize enterprise executives, those same execs are insisting that IT somehow make it happen.
When Google rolled out its latest biometrics specs for Android devices, its top-level 'strong security' option allowed “a spoof and imposter acceptance rate not higher than 7%.” Most biometrics specialists argue that's muc....
The idea that vendors lie a lot is, as the saying goes, “a tale as old as time.” But to suggest vendors are so persuasive because they actually believe their falsehoods — now, that's intriguing.
Zoom stirred up a kerfuffle this month when it amended its terms of service to make execs comfortable that it wouldn’t use Zoom data to train generative AI models. In reality, it was really doing spin control worthy of the sleaziest politician.....
As details about the recent China attack against US government agencies come to light, two details stand out: Microsoft failed to store security keys properly — and the keys were used by attackers even though they'd already expired.
In many ways, lawyers, CIOs and CISOs have the same mission: protect the enterprise from forces that want to do harm. But those two professions often approach the task in such polar opposite ways that they fight each other instead of the bad guys.
Gaining visibility into anything IT-related is always difficult, but the age-old nemesis, shadow IT, remains a major problem — especially as the enterprise environment has changed.
If coders lied as often as ChatGPT, they would be fired immediately. Stunningly, some enterprise execs seem to be just fine with that — as long as AI continues to code quickly and for so little money.
Using generative AI to code is dangerous for a variety of reasons, but its efficiencies will tempt corporate leaders — especially CIOs and business execs — to use it anyway. A senior AWS executive at Amazon argues the decision doesn&rsquo....
As we’ve seen with other highly-hyped technologies — such as the Web back in ‘95 and blockchain more recently — companies can get ahead of themselves when they jump into investments based on things other than strategic goals. ....
A compliance fight between Microsoft and German regulatory authorities has gotten white hot, though it looks as though any penalties might bypass the company and take aim at its customers.
Biometrics are supposed to be a fundamental pillar of modern authentication. Unfortunately, for a wide range of reasons and in a variety of ways, many biometric implementations are wildly inaccurate.
As the COVID-19 pandemic slowly fades — and the rush to cloud solutions it hastened now seems less critical to business success — a question arises: Has anyone on your team recently run an ROI analysis to see whether the cloud truly save....
This goes beyond simply not trusting location data for cybersecurity authentication. Geolocation is now used for a wide range of business reasons — but it shouldn’t be.
The US Federal Reserve and the US Senate are both looking to lessen restrictions on retailers — ostensibly to rein in card fees. What they actually are doing is inviting more fraud.
The European Union is cracking down on cryptocurrencies. That could have massive implications for enterprise IT.
Microsoft is backing off its support for some AI-driven features, including facial recognition. Although it's good Microsoft is acknowledging discrimination and accuracy issues, it had years to fix the problems and didn’t.
There are disturbing reports that some major financial institutions are no longer crediting back all fraudulent transactions, even when the victim has filed a police report. This move by these financial institutions will soon come back to bite them. ....
C-level execs argue a fine game about caring about their employees — but those platitudes somehow never make it into the HR meetings about bonus benchmarks.
What if smartphone sound-recognition could be tweaked to do core IT and operational chores? This would be an option to customize the phone to listen for sounds specific to your company.
It’s not often that you see two cybersecurity vendor CEOs agree on an issue — and yet get into a very public insult-fest with each other. Then again, this did start at RSA, so anything is possible.
A June report from an analytics firm has Amazon knocking Walmart out of its No. 1 retailer slot by 2024. Walmart bet on a store-based approach years ago, but consumers changed their habits and Walmart is soon to pay the price.
Given that one of the uglier threats to enterprise cybersecurity involves re-purposed third-party code and open-source code, you might think that Google addressing the issue would be a big help. Think again.
Sponsored Links