By Adam Mills
AI Editor
0
Share
Newsweek is a Trust Project member
See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.
Newsweek
Editor's note: This is AI Impact, Newsweek's weekly newsletter where each week, we will explore how business leaders are unlocking real value through artificial intelligence.
Read More on AI
Tap or click here to get this newsletter delivered to your inbox.
Signal Capture
Signals from the frontlines of AI adoption
How AI Raised the Bar for Lawyers
By Adam Mills
Legal departments were once known for slowing contracts down. AI is now changing that reputation, though not in the way many lawyers may have expected.
Ironclad, an AI contracting platform, found in its third annual “State of AI in Legal” report that 92 percent of legal professionals now use AI in some capacity for legal work, up from 69 percent in 2025. The survey included 822 legal professionals, evenly split between in-house legal teams and law firms.
Among AI users, 88 percent said their workload has increased since adopting the technology. The report also found that 89 percent are spending more time on complex, strategic tasks.
Singh said the findings show why the workload question is more complicated than it first appears.
“AI is not just reducing workload; it is changing what legal teams are expected to deliver,” Jasmine Singh, general counsel at Ironclad, told Newsweek. “When you make the first pass faster, you do not necessarily end up with less to do. You end up with more capacity to engage in higher-value work that was previously getting deferred, rushed, or not done at all.”
Ironclad
Much of that adoption is happening inside contract work. Ironclad found that 94 percent of respondents who use AI for legal work use it for contract-specific tasks.
“Contracting is where legal AI has shown the clearest value because much of the work is structured, repeatable and tied to known standards,” Singh said. “Legal teams already have playbooks, fallback positions, risk thresholds and recurring issue patterns, which makes contracting a particularly good fit for AI-assisted first-pass work.”
The payoff is becoming easier to measure. Among legal professionals using AI, 97 percent reported at least one business outcome, including faster response times to stakeholders, faster contract turnaround and reduced outside counsel spend, according to Ironclad.
The report also found that 96 percent of respondents said their organization now expects more from the legal function than it did two years ago.
“Historically, many teams viewed legal primarily as a control function: review the contract, flag the risk and sign off,” Singh said. “That is no longer enough.”
As more first-pass work moves through AI systems, the remaining questions become harder to automate.
“This is the distinction I often make between rules and judgment,” Singh said. “AI is well suited to the rules side of legal work: first-pass review, issue spotting, clause comparison, summarization and other repeatable analysis. Lawyers still need to own the judgment side: when to flex, when to hold the line, how to weigh business context and how to navigate situations where the answer depends on leverage, risk tolerance or the human dynamics of the deal.”
Legal’s AI use is also drawing attention from outside the department. Ironclad found that 98 percent of legal professionals have been asked by people outside the legal function about legal’s approach to AI usage.
“Legal teams have become a leading example of how companies can adopt AI responsibly while maintaining accountability,” Singh said. “Legal teams are adopting AI in high-stakes environments that require careful oversight, so other departments are turning to legal for guidance on AI usage and compliance.”
That visibility does not resolve the hardest question around legal AI: What happens when the technology is wrong?
Ironclad found that 96 percent of respondents said they would use AI more extensively in legal work if responsibility for errors were more clearly defined. Singh said errors can originate in the user, prompt, model, tool configuration, approval process or review process.
“But complexity does not eliminate responsibility,” Singh said. “In legal especially, AI cannot become a way to blur ownership. Lawyers still need to review, exercise judgment, and stand behind the final output.”
Core Intelligence
AI’s Hidden Productivity Tax
...
By Adam Mills
AI productivity math looks increasingly incomplete.
A new GoTo and Workplace Intelligence study found that employees using AI save an average of 2.3 hours a day, a figure that would seem to validate the corporate push to bring generative tools into everyday work. But the same research suggests some of those gains may come with hidden costs companies are not yet measuring.
“We’ve reached an inflection point with AI adoption,” Rich Veldran, CEO of GoTo, a business communications and IT software company, told Newsweek. “A year ago, everyone was using AI, but they weren’t quite getting real productivity gains. That’s changed now.”
The challenge for employers is that faster work can still become more expensive if it weakens judgment, creates rework or pushes cleanup onto colleagues asked to fix AI-generated drafts.
“The productivity gains are real and measurable, but there’s real concern that when not used properly, AI is also eroding skills, judgment and accountability,” Veldran said. “Companies need to understand that productivity and capability are not the same thing. You can get faster without getting better.”
Leaders may need to change how they measure AI’s value. A worker who finishes a document faster is not necessarily producing work that requires less review. A company that counts hours saved may still miss the time lost when employees have to verify, repair or redo AI-assisted work.
Dan Schawbel, managing partner of Workplace Intelligence, a workplace research and thought leadership firm, said the cost goes beyond time.
“Productivity gains are real, but they’re arriving alongside a quieter crisis of confidence that most leaders aren’t tracking,” Schawbel told Newsweek. “The gap between AI’s visible outputs and its invisible psychological costs is the story organizations are missing entirely.”
The problem is especially sharp for early-career workers. According to the report, 46 percent of Gen Z employees say AI is weakening their skills and making them less intelligent, while 50 percent believe relying on AI will hurt their long-term career prospects.
“If AI is short-circuiting that development process, we’re creating a generation of workers who may struggle with the higher-order thinking that AI can’t replicate,” Veldran said.
The accounting problem also shows up when AI-generated work lands with someone else. The study found that 59 percent of employees are responsible for reviewing AI-generated content created by coworkers or direct reports. Among those reviewers, 77 percent say AI-generated work takes more time to review than human work, while 66 percent say reviewing other people’s AI “workslop” creates additional work for them.
Individual speed can become a team-level tax.
“If AI helps you create a document in five minutes instead of an hour, but then it takes your colleague 45 minutes to review and fix it instead of 15, where’s the net gain?” Veldran said. “Leaders need to think about productivity at the system level, not just the individual level.”
AI can create the appearance of progress before the organization knows whether the work has improved. Companies need to measure what happens after the output appears: whether quality rises, whether rework falls, whether employees trust the work and whether responsibility for errors is clear.
Without that discipline, AI adoption can turn into a workplace shell game. The hours saved by one employee can reappear as review work for someone else, while higher output masks declining confidence in the work itself.
Schawbel warned that companies could automate their way into a weaker workforce if they treat deployment as the finish line.
“Employers need to treat skill development with the same urgency they brought to AI deployment,” he said, “or they’ll find themselves with highly automated organizations full of people who’ve forgotten how to think without a machine.”
You can read the article here: AI Is Making Workers Faster. That May Be the Problem.
Upcoming Webinars
AI in Finance: From Individual Adoption to Enterprise Transformation
Financial services firms are beginning to face a harder AI question than whether the technology can simply make professionals more productive. The real test is what they do with those gains.
In an upcoming “AI Impact Forum” session, Dr. Ranjit Tinaikar speaks with Kevin Buehler, chief innovation officer at Rogo and senior partner emeritus, McKinsey & Company, about how agentic AI could reshape financial services—and whether productivity gains become cost savings, new ways of working or a lasting competitive advantage.
Join the live discussion on Thursday, June 18, at 9:30 a.m. Eastern. Register, for free, right now.
Prompt Injection
What’s one recent insight you’ve learned about AI?
...
“The rapid rise of AI is, paradoxically, making the internet more human, at least in terms of how most of us experience it. The internet was already flooded with too much content, and AI slop is exacerbating that.
AI search is already a more human experience than Google. When I ask a search engine a question, I get 10 blue links, seven of which are paid placements. Yuck. When I ask ChatGPT or Claude the same question, I get a ‘thoughtful’ answer. Maybe that answer isn't great, but it's a much more human-feeling interaction. This is why AI search will win and blue links will disappear over the next few years.
For people, this is great! We get better answers.
For companies, this is hard. For 25 years, digital marketers have been able to follow a formula and have success. Now, that formula doesn't work as well. For businesses, the marketing pendulum is swinging away from science and back towards art. Not art in a museum, but art in a sense that we have to strike an emotional chord (not rage!) to gain attention and build trust. It's a big mindset shift that demands confidence and daring.”
Have your own lesson to share? Email us at: a.mills@newsweek.com
Run Log
AI use case of the week
...
A disputed health care claim can splinter quickly. One answer might come through a provider portal, another through email and a third from a staff member working from a different record. In the Independent Dispute Resolution (IDR) process, where payers must respond within legal timelines, those gaps can turn already complex casework into a volume problem.
Daffodil Health, a company that uses AI to automate health care pricing operations, is trying to keep those disputes from fragmenting. It uses AI agents that draw on each payer’s policies, fee schedules, contracts and claim histories to manage high-volume interactions with providers, employers and members.
The key is making sure every channel is working from the same case file. Each interaction reads from and updates that record, so context can carry across channels, over time and when a case moves between AI and human staff. That reduces the risk of a provider getting one answer from a chatbot, another from email and a third from someone reviewing incomplete information.
Shuo Yang, chief technology officer and co-founder of Daffodil Health, said the goal is not to replace claims staff. It is to automate routine communication so people can focus on escalations, complex negotiations and exceptions.
Early work, Yang said, has shown why the shared case file matters. When AI and human teams work from the same record, responses are less likely to contradict each other, teams can move faster within legal IDR windows and payers can handle more disputes without adding staff every time volume rises.
“The market doesn’t lack AI tools — it lacks AI systems,” Yang said.
Have an AI use case to share with us? Email us at: a.mills@newsweek.com
Context Window
■ Recent college graduates are facing a tougher job market than the broader workforce, with economists divided over whether AI is hurting entry-level hiring in exposed fields or whether a broader slowdown in new job creation is the bigger factor. [ABC News]
■ Airbus has signed a partnership with Mistral AI to deploy trusted AI across aerospace operations, engineering, onboard systems and defense applications, with on-premises and trusted cloud options for sensitive industrial and military use cases. [Airbus]
■ Illinois lawmakers passed an AI accountability bill modeled on laws in New York and California, a further sign of state-level pressure on large model developers to disclose safety frameworks, use third-party auditors and assess catastrophic risks. [25News Now]
■ Cisco researchers found that leading AI models were far more vulnerable to multi-turn attacks than single-prompt safety benchmarks suggest, a security and governance risk for CISOs using public model scores to guide deployment decisions. [CSO]
■ Salesforce used record quarterly revenue and growth in Agentforce and Data 360 to push back against fears that AI coding agents will weaken SaaS incumbents, though Wall Street remained cautious after a softer-than-expected revenue forecast. [The Daily Upside]
■ Atlassian has put AI enablement under its chief people officer, a sign that the company sees enterprise AI adoption as a culture, workflow and employee-trust challenge as much as a tooling issue. [Unleash]
Transfer Protocol
Tracking executive moves across the AI landscape
Eric Curwin, previously vice president of data and AI at AlertMedia, has been promoted to the company’s first chief AI officer, where he will lead AI design and development, set architectural standards and guide responsible adoption across the platform and business.
Justin Bass, an engineering, AI implementation and enterprise technology transformation leader, has joined Investor Group Services as chief AI officer, co-leading the firm’s AI practice and helping private equity clients develop, prioritize and execute AI strategies.
Sam Dozor, a founding engineer at mParticle who spent more than 13 years helping build the company into a leading customer data platform, has been appointed chief technology officer at Rokt, taking charge of engineering, platform infrastructure, developer experience and security as the company builds its AI-driven software platform.
Saugata Saha, previously president of S&P Global Market Intelligence and chief enterprise data officer at S&P Global, will become president and chief executive officer at Qlik, with a mandate to lead the data, analytics and AI company’s next phase of product innovation and commercial growth.
Zain Lakhani, founder of the AI engineering firm LightCI, has joined Pendo as its first chief AI officer, where he will oversee AI engineering and help accelerate the development and deployment of production-grade AI systems across architecture, product development and partnerships.
Know someone on the move in AI? Send job change info to a.mills@newsweek.com
Magic Moment
What’s the most fun or unexpected way you’ve used AI lately?
...
“The most unexpected way our team has used AI lately has nothing to do with real estate. We use it to make absurd images of animals as historical figures and send them to each other on Slack. A capybara as Napoleon. A golden retriever as Lincoln. A pigeon, somehow, as Genghis Khan.
It started as a one-off joke and it has become a thing. We have a channel for it now. Specifically, we got obsessed with slug images. People drop one in when a deal closes, when somebody’s having a rough Monday, when there’s nothing in particular going on at all.
What surprised me was what it actually built. We’re a small team moving fast on a hard product. The work is real and the stakes are real. The slugs are not. But the camaraderie of a group chat that makes each other laugh every day is something I would have paid a consultant to engineer. Instead, it showed up sideways, through a tool everyone is supposed to be using for productivity.
AI made us more efficient at work. It also made us friends.”
Experience some AI magic? Tell us about it at a.mills@newsweek.com
Tap or click here to get this newsletter delivered to your inbox.
View 0 Comments
Submit Correction
View Editorial & AI Guidelines