IT STRATEGY

Francis Scialabba

AI tools aren’t perfect for every job—especially when it comes to hiring staff.

Recruiters are increasingly turning to automation tools to pull prospects from an ever-growing pile of résumés. While that has sped up the hiring process, it has also introduced a fair degree of uncertainty: a 2026 report from hiring platform Greenhouse of 1,865 US respondents (including 1,200 job-seekers and 665 recruiters and hiring managers) revealed that just 21% of recruiters are “very confident” their systems aren’t rejecting qualified candidates.

There are also persistent fears that AI is leading to biased outcomes in hiring. In another study, researchers from Stanford University, Chapman University, and Northeastern University analyzed 4 million job applications and found that automated assessment produced unequal outcomes across racial groups.

If that wasn’t bad enough, those researchers also found that “people who submit multiple applications to positions screened by the same algorithmic hiring vendor are more likely to be rejected from every position to which they apply than would be true if the companies made decisions statistically independently from one another.” For IT pros hunting for work, in other words, algorithms can work against you if you’re applying to jobs at multiple companies.

Hiring managers share their AI strategies.—BH

      

Sponsored By PwC

Paradox alert: While token costs may be dropping in some cases, AI bills are still surging. Wait, what?

PwC shares that seeming “cheapness” makes volume skyrocket. Everyone wants to use AI everywhere, even if the result of this is more complex, expensive workflows. And that’s the key problem: using AI everywhere rather than where it really makes a difference.

In this article, PwC explains how CIOs and CFOs can scale AI with discipline to achieve cost efficiency.

PwC notes that AI value per cost is pretty crucial for the C-suite to understand. It may soon be what allows you to transform operations, enter new markets, and stand up new business models, while competitors may still scramble to find the funds. Read more.

CYBERSECURITY

Getty Images

A shortcut is great—unless it sends you off a cliff. That applies both to cross-country road trips and AI coding assistants used to generate software.

In a July blog posting, Maor Dokhanian, a threat researcher at cloud and AI security tools provider Wiz, discussed the discovery of a systematic vulnerability, dubbed GhostApproval, that occurred in six of the top AI coding assistants.

The affected coding assistants included Amazon Q Developer, Anthropic Claude Code, Augment, Cursor, Google Antigravity, and Windsurf. The vulnerabilities could allow malicious actors to trick the agent into “accessing arbitrary files outside the workspace sandbox, potentially achieving remote code execution on the developer’s machine,” Dokhanian wrote.

One of the steps includes a tricky README file.—CN

      

SOFTWARE

Getty Images

As AI increases its capabilities, there’s less of a need for users to give their chatbots lengthy directives—which is leading to some interesting changes, one company found.

Research from Elvex, an enterprise AI adoption platform, indicates that AI prompts are shortening as the underlying technology gets better at predicting what users want. That’s leading to higher positive feedback from customers, Elvex CEO Sachin Kamdar told IT Brew.

“They’re getting more back from the systems because of integrations and context, and they’re liking those responses more,” Kamdar said.

Kamdar acknowledged that Elvex is relying on internal data, but he suspects there’s a universality to the findings once you control for customer quality.

The prompt evolution is changing hiring, too.—EH

      

patch notes

Francis Scialabba

Today’s top IT reads.

Stat: 62%. That’s the proportion of ransomware victims holding the title of manager or higher, according to recent research from cybersecurity platform Zscaler’s ThreatLabz intelligence unit. The group studied 351 victims across 334 organizations, all linked to a single ransomware campaign. (ZDNet)

Quote: “I created an accidental honeypot.”—Cory Solovewicz, after buying noreply.net and noreply.us domains and receiving a deluge of inadvertently shared emails and attachments (Ars Technica/Wired)

Read: In the age of AI, some vintage computer collectors are preserving classic machines. (theAssociated Press)

^{*A message from our sponsor.}

Getty Images

Learn how shorter LLM responses can help IT teams reduce AI costs without sacrificing results.