When I first heard about OpenClaw, it sounded amazing. Or, at least I thought it sounded amazing. An AI agent that will run your personal life, keeping your inbox and calendar in line? Acting on my behalf overnight, doing the research I need even when I’m not asking for it? But my inner skeptic kicked in pretty loudly: Can I actually trust an AI agent to do all of that?

Security professionals weighed in pretty quickly: You shouldn’t without some meaningful guardrails—and even then, given all of the power the AI agent will have, you might want to think twice. Reddit has also been full of cautionary tales from people who should know better about OpenClaw agents going rogue, deleting data, introducing malware and doing things the user doesn’t want. A new study from enterprise cybersecurity provider Ivanti finds that 77% of cybersecurity professionals are willing to trust AI agents to make decisions without their input, so this isn’t necessarily surprising.

I spoke with Ivanti Chief Legal Counsel Brooke Johnson about it, and she said that CIOs—and everyone else—need to remember that AI agents’ actions have the same legal weight as those of people, which means they carry real liabilities. I talked to her about what CIOs need to consider, and an excerpt from our conversation is later in this newsletter.

Until next time.

This is the published version of Forbes’ CIO newsletter, which offers the latest news for chief innovation officers and other technology-focused leaders. Click here to get it delivered to your inbox every Thursday.POLICY + REGULATIONS

President Donald Trump speaks at the “Winning the AI Race” summit hosted by All‑In Podcast and Hill & Valley Forum in Washington, D.C. last July.

Chip Somodevilla/Getty Images

The Trump administration has backtracked a bit on its stance against any AI regulations. This week, the Commerce Department announced agreements with Google, Microsoft and xAI to allow the government to review new AI models before they are publicly released. The reviews will be done by the Center for AI Standards and Innovation, an agency under Commerce’s National Institute of Standards and Technology.

It’s not clear what these reviews will entail or if the government will require companies to make changes if issues are found. Frontier AI systems have had several well-publicized issues, ranging from OpenAI’s chatbots giving weapon-making instructions to Anthropic’s very tightly controlled Mythos platform, which can find critical cybersecurity vulnerabilities in seconds. In a blog post, Microsoft writes that it expects to be working with CAISI to test AI systems in ways that probe unexpected behaviors, like automobile crash tests.

It’s also unclear if politics will come into this process. Following a high-profile dispute with Anthropic over a Pentagon contract—in which the government rejected limits on how the AI platform could be used—Anthropic has been mostly excised from government use and deemed a supply chain risk. Last week, the Pentagon announced agreements with seven other tech firms to use AI for classified military networks: most of the big players except Anthropic. And the head of CAISI, who will be running the AI systems testing, was just nominated last week; the previous director was asked to resign because he had previously worked for Anthropic, the Washington Post reported.

FROM THE HEADLINES

AI is once again a top reason companies cite for cutting jobs, according to new data from career services firm Challenger, Gray and Christmas. In April, the firm found 83,387 jobs were eliminated, a 38% increase from March. And companies said more than a quarter of those were because of AI—but experts are skeptical.

Forrester analyst J.P. Gownder told Forbes senior contributor Caroline Castrillon that the vast majority of companies are a long way from building the AI infrastructure to actually replace that many people. These layoffs are more likely companies “right-sizing” their workforces following the Covid-era buildup of employees. AI will definitely have a huge impact on the workforce—Forrester predicts 6.1% of all jobs will be affected by 2030—but given where the technology is today, it’s not quite so sweeping yet. Forbes contributor Bernard Marr writes that the biggest impact will be at the bottom of the career ladder—entry-level jobs—and not among today’s professionals.

Meta founder and CEO Mark Zuckerberg’s reasons for a planned cut of about 8,000 employees—10% of Meta’s workforce—might actually be a little more about AI than most. The Wall Street Journal reported Zuckerberg said at an internal meeting last week that AI compute costs are a major reason for layoffs. Infrastructure costs have played an outsized role on Wall Street for Meta and other tech companies. Meta beat expectations and grew revenue in its earnings last week, but its share price dropped more than 8% based on an $10 billion increase in infrastructure costs this year, writes Forbes senior contributor Peter Cohan. Forbes’ Alicia Park writes that Big Tech is slated to spend $750 billion on AI infrastructure this year.

BITS AND BYTESThe Real Liability Of AI Agents

Ivanti Chief Legal Counsel Brooke Johnson.

Ivanti

AI agents are an exciting new technology for the enterprise, with 87% of cybersecurity professionals telling enterprise IT security company Ivanti that adding it is a priority. But they may be placing too much trust in them, with 77% saying they’re comfortable with AI agents taking actions without human approval.

Ivanti Chief Legal Counsel Brooke Johnson agrees AI agents can be a game-changer, but they can carry real legal and liability risks for companies that CIOs need to be aware of. I talked to her about what CIOs need to know to protect their companies. This conversation has been edited for length, clarity and continuity.

If your AI agent does something wrong, who is responsible from a legal standpoint?

Johnson: I believe employers and companies are responsible for the actions of their agents. It’s the same concept of vicarious liability for an employee’s actions when they’re engaged in work on behalf of the company. It has to be the same thing for agents. If an individual employee is in charge of an agent and the agent does something, the individual employee is not the one who’s going to ultimately have the deep pocket to remedy the issue. Liability will flow back to companies.

From a business perspective, we’re contracting for the shared risk. That’s a very normal part of software: we agree to split liability in different ways. That will continue to evolve to make sure that agents are included, then businesses are prepared to hold the risk associated with agent activity. We have insurance policies and general practices.

Companies also wouldn’t hire an employee and not train them. You’re responsible to train them. You’re responsible to train your AI, to oversee it, to make sure you have the controls in place internally so that it is carrying out the actions you want it to.

Do most companies realize and understand this? Are they ready to take the liability if an agent goes rogue, or is this being overlooked by CEOs?

It’s an evolving area of business and the law. I can’t think of any specific lawsuits around agent activity, but I think it’s something that any C-suite should be thinking about, and I suspect we will see cases around this.

If you’re getting in front of it, you should go back to what are your governance procedures: How are you training your human employees about what’s allowed and what’s not, and then providing them the right paths to getting tools approved that really will help the business overall.

I think insurance companies will probably drive some of the conversation around it because your cyber liability policy may start specifically speaking about agents. It might evolve that way: What they’re willing to cover, what they’re not, if there’s any controls you have to have in place in order to receive coverage.

The courts are probably going to be out in front of U.S. regulators, influencing how businesses manage risk and potential liability. The U.S. system often is driven from the ground by litigation because there is harm and injury. People seek the remedy in courts through monetary awards—which always drive business decisions and the management of risk.

It’s happening so fast, though. I read a New York Times article about OpenClaw in China. It’s driven a tech boom, and then the Chinese come in and shut it down because they’re so concerned about the risks of it. It really is this interesting wave of things happening: The technology is out there, then people are having to react after the fact to what’s going on.

What advice would you give to a CIO to make sure that they’re on the right track when it comes to potential legal issues with agents?

Talk to whoever it is at your company that runs data. One of the key pieces that helps keep you out of trouble is figuring out what data is it going to have access to, what are the risks associated with that, and what are the regulatory requirements around giving it access to that data? Work with your head of security. They can help walk through what is the impact to the corporate environment and what are the access points. They can help build that structure [similar to what you’d want for] managing an irresponsible employee: How do you create the right infrastructure to responsibly be able to manage and review?Think about audits. Is there oversight and tracking, and are we able to go back and see if it did the wrong thing? Do we have logging? Do we have the right roadmap for what it’s been doing so that we can run an investigation?COMINGS + GOINGSOmnichannel commerce platform Lightspeed Commerce selected Bhawna Singh to be its new chief technology officer. Singh most recently worked as the chief technology officer at Okta, and she held the same role at Glassdoor prior to that.Aircraft leasing firm Aircastle appointed Nadene McKenzie-Reid as its new chief information officer, effective June 1. McKenzie-Reid joins the company from NatWest, where she worked as the head of technology.Herbal and nutritional supplement manufacturer Nature’s Sunshine hired John Hnanicek as its new chief technology officer, effective May 4. Hnanicek steps into the role from Melaleuca, where he worked as chief digital innovation officer.STRATEGIES + ADVICE

When you make the decision to use a chatbot instead of a human helper, it makes sense to try to make it seem friendly, but researchers have found that might not be the best move. “Friendly” chatbots are more likely to make factual errors in efforts to be agreeable—and users find them more annoying.

One potential use of AI you may have overlooked: Personal brand building. Here are 11 ways to use it to help you stand out professionally.

QUIZ

To warn people about AI deepfakes, an influential woman reposted an AI-generated image of herself on a bed in lacy lingerie on social media. Who is it?

A. Rep. Nancy Pelosi

B. European Commission President Ursula von der Leyen

C. AMD CEO Lisa Su

D. Italian Prime Minister Giorgia Meloni

See if you got the answer right here.