AI is supposed to make work easier, but instead it has generated a new problem: "workslop."

Why it matters: The term, coined by researchers in the latest Harvard Business Review, describes low-quality AI-generated content — memos, reports, emails — that's clogging up employees' lives and wasting their time.

What they're saying: Workslop "appears polished but lacks real substance," write researchers from Stanford University who collaborated with BetterUp, a leadership coaching platform.

• "[Y]ou might recall the feeling of confusion after opening such a document, followed by frustration," they write. "You begin to wonder if the sender simply used AI to generate large blocks of text instead of thinking it through."
• "If this sounds familiar, you have been workslopped."

By the numbers: In August and September, the researchers surveyed 1,150 U.S. adults who described themselves as desk workers about their experiences with "workslop." They didn't name the term, but merely defined it.

• 40% said they'd encountered this stuff in the last month — slowing down their workday.
• They reported spending an average of 1 hour and 56 minutes dealing with each instance.

Zoom out: Slop carries real costs — looking at the average respondent salary, the researchers estimate workslop incidents costs $186 per month.

• For a large organization, that can add up to more than $9 million a year in lost productivity, per their back-of-the-envelope math.
• When asked how workslop made them feel, more than half of respondents said they were annoyed, 38% were confused and 22% were offended.

The intrigue: Colleagues look down on workslop senders — about half of all those surveyed viewed slop senders as less creative, capable and reliable.

• And yet they're sending it, too. Of those using AI at work, 18% admitted sending AI-generated content that was "unhelpful, low effort or low quality."

The big picture: Workslop is the workplace offshoot of the general run of AI-generated slop most of us see day-to-day — rabbits jumping on trampolines, fast fashion ads featuring Luigi Mangione, weird uncanny images of hands with the wrong number of fingers, and so forth.

• It's also another sign that AI isn't necessarily translating into productivity gains at work.
• A report from MIT out a few months ago found that 95% of business' AI pilot projects fail.

Zoom in: The researchers noticed workslop in their daily lives, as friends, colleagues and families shared frustrating experiences, said Jeffrey Hancock, director of Stanford's Social Media Lab.

• Survey respondents also shared examples — including from health care providers who griped about getting long AI-generated reports from patients that diagnose their health problems using data from Fitbits or Oura rings, without any real medical underpinning.

Yes, but: Well before the advent of AI, employees were generating poorly constructed memos, PowerPoints and emails.

• Researchers told Axios the workers reporting the most slop were in tech, health care and professional services. (Consulting is basically ground zero for the overwrought slide deck.)

The bottom line: You can use AI to make your work better, says Kate Niederhoffer, vice president of BetterUp Labs and one of the researchers.

• But you can also use it to pretend to get 20 tasks done, she says, and just "Trojan horse" a bunch of work to your colleagues.

A host of CEOs will launch an ad campaign today to push companies, governments and other organizations to speed up their efforts to adopt AI.

Why it matters: Governments around the world are looking for guidance on how to build AI-driven economies, and the Business Software Alliance — the industry group behind the campaign — is looking to influence policymakers' approach in the U.S. and abroad.

• "In this era of global competitiveness, the AI advantage stands to transform every business and government agency who recognize the value of modernizing stagnant processes like contracting with practical AI solutions," Docusign CEO Allan Thygesen says in an ad kicking off the series of testimonials.

Driving the news: The CEOs of Autodesk, Bentley Systems, Cisco, Cohesity, PagerDuty, Rubrik and Siemens will also be featured in the first wave of the campaign.

• CEO testimonials will be shared with elected officials, policymakers and the media, and the videos will be promoted with a paid campaign on LinkedIn targeting those same audiences.
• For example, Thygesen says in his testimonial that Docusign has seen the value of the AI it has already deployed: "AI makes contracts searchable, turning them into a business or a government's most valuable asset — a source of insights and efficiency."
• The ad campaign will run at least through early November.

Context: BSA, the leading global trade group for the software industry, organized and paid for the ad campaign.

• The trade association's pitch to policymakers for AI adoption in the U.S. includes calling for workforce training, expediting permitting for AI and energy infrastructure, and for Congress to pass AI legislation that establishes a single national standard, among other proposals.
• BSA declined to disclose the ad buy figure.

OpenAI's Stargate data center infrastructure project will add five new sites to provide seven gigawatts of planned capacity over the next three years, the company announced yesterday.

Why it matters: The push to build massive, power-thirsty data centers — which both tech giants and government leaders see as essential for the U.S. to dominate global AI — is entering a frenetic new phase.

Between the lines: OpenAI made the announcement at its flagship site in Abilene, Texas, in partnership with Oracle and SoftBank, alongside Sen. Ted Cruz (R-Texas) and Rep. Jodey Arrington (R-Texas).

• The additional sites will be located in Lordstown, Ohio; Shackelford County, Texas; Milam County, Texas; Doña Ana County, New Mexico; and an additional site in the Midwest to be named later.
• OpenAI said in a blog post that the company is investigating additional sites.

The big picture: OpenAI says the new sites put the firm ahead of schedule to lock in its previously announced $500 billion, 10-gigawatt commitment by the end of 2025.

• SoftBank CFO Yoshimitsu Goto admitted last month that the Stargate project faced issues with site selection and coordinating stakeholders, putting the project behind schedule.
• But OpenAI now says Stargate has $400 billion of the originally announced $500 billion goal lined up over the next three years.

What they're saying: "AI can only fulfill its promise if we build the compute to power it," OpenAI CEO Sam Altman said in a company blog post today.

• In a separate post on his personal website Altman warned that limiting compute would force society to make choices between competing priorities — like curing cancer or providing a personalized tutor for "every student on earth."
• "If we are limited by compute, we'll have to choose which one to prioritize; no one wants to make that choice, so let's go build," Altman wrote.

Go deeper: How is Stargate's $500B getting funded?

• Scoop: Microsoft is in talks with U.S. publishers to create a pilot marketplace for AI makers to license their content for use in AI services. (Axios)
• Al Gore is warming up to nuclear power in the age of AI, he told Axios in an exclusive interview.
• The developers behind Google's hit NotebookLM AI research tool and podcast maker raised $4.6 million to launch a new personalized audio news service called Huxe. (TechCrunch)

Here's an easy way to determine if that recruiter on LinkedIn is a bot or not.