IT OPERATIONS

Illustration: Morning Brew Inc., Photos: Adobe Stock

Halloween may still be a month away, but something frightening may be haunting your business in the meantime: ghost assets.

Ghost assets cannot be accounted for, but still show up on a company’s inventory list. For example, they might include devices that have been lost, misplaced, or stolen.

“They could be left in a desk drawer. They could have left with an employee who took it with them when they departed the organization,” Mitch Berk, VP of product management and autonomous endpoint management at Ivanti, said. “They could have been recycled and not deleted from the systems, but it’s an accounting or systems error where things aren’t reconciled.”

Henrique Teixeira, SVP of strategy at Saviynt, told IT Brew ghost assets can also include digital assets, including accounts and identities.

Welcome to your own haunted mansion.—BM

      

Sponsored By HUMAN Security

Customers just don’t shop like they used to. That’s because AI assistants are doing it for them: discovering, comparing, recommending, and (increasingly) transacting on their behalf.

According to HUMAN telemetry, agentic traffic has surged more than 6,900% (cue the double take). Adobe research also shows:

• 38% of consumers have used GenAI for shopping
• 85% said agentic shopping improved the experience

Visitors arriving through AI-generated answers convert 4.4× better than visitors from traditional search. To be surfaced and trusted, brands need accurate, consistent, machine-readable catalogs, policies, feeds, and APIs—plus infrastructure that can support machine-speed interactions.

This guide’s nine-step readiness framework recommends governed participation. What that means: Allow trusted agents to browse and transact, throttle excessive behavior, and stop harmful automation in real time.

Read the guide here.

IT STRATEGY

Kilito Chan/Getty Images

AI is evolving faster than you can say “Jack Robinson,” and some AI leaders want to change that.

On Sept. 12, Anthropic co-founder and CEO Dario Amodei penned an essay calling for the broader AI industry to slow its pace of development. In his essay, Amodei shared a “three-step plan” to do so, which recommends:

• Embedding third-party evaluators (e.g., METR) within leading frontier organizations to report safety incidents and verify safety commitments are being met
• Collaboration among frontier AI companies within democratic countries “to establish common safety standards as well as limits on the rate of unchecked AI progress”
• Cooperation with authoritarian governments to achieve global regulation

“I believe that if slowing down bought us even an extra year or two before models reach critical levels of capability, and we used that time to advance alignment, we could greatly reduce the risk that something goes seriously wrong,” Amodei wrote.

Slow down, everyone.—BM

      

CYBERSECURITY

Niv Bavarsky

AI risk is all over the news as the public tries to understand the threat that models and agents pose. But for IT teams, the real question is how much the danger changes their security posture, and how to communicate that change to their organizations.

Recent, well-publicized breaches like the Hugging Face attack have highlighted the importance of managing the risk of increasingly capable AI agents. Whether that marks a new risk or an increase in a known one remains an open question.

Changing environment. Ramy Rahman, senior solutions engineer at ArmorCode, told IT Brew that he sees the potential for damage as real and worth planning for. The new agentic swarms are fundamentally different from older bots that lacked the capabilities of modern AI.

Communicating that change in power and capability outside of the IT team, though, can be daunting.

Thinking through the threat.—EH

      

patch notes

Francis Scialabba

Today’s top IT reads.

Stat: $6 trillion. That’s the annual AI revenue needed to justify data center infrastructure buildout, according to Bain & Company. (Network World)

Quote: “An agent doesn’t need malicious intent to create a security incident. It needs enough access and one bad decision.”—Ryan McCurdy, vice president of database change-management company Liquibase, on the need to control AI agent access (Dark Reading)

Read: Meta’s Muse AI assistant is creating in-depth profiles for everyone in a user’s life. (Wired)

^{*A message from our sponsor.}

Jobs

CollabWORK connects you to the hidden job market through IT Brew and other trusted channels. Browse roles curated specifically for this community by clicking through to the job board.