The AI Daily Brief: Artificial Intelligence News and Analysis - Workers Don't Trust Their Companies on AI

Episode Date: August 30, 2025

Stanford research shows AI is already reshaping entry-level jobs, with early-career workers in high AI-exposure fields seeing a 13% drop in employment. At the same time, a survey by Kyla Scanlon finds most employees don’t trust their employers’ AI strategies—with over a third expressing zero trust. Despite this distrust, few companies are offering AI training, creating a dangerous gap that could fuel resistance and sabotage of corporate AI initiatives.Brought to you by:KPMG – Discover how AI is transforming possibility into reality. Tune into the new KPMG 'You Can with AI' podcast and unlock insights that will inform smarter decisions inside your enterprise. Listen now and start shaping your future with every episode. ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.kpmg.us/AIpodcasts⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Blitzy.com - Go to ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://blitzy.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠ to build enterprise software in days, not months Vanta - Simplify compliance - ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://vanta.com/nlw⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Plumb - The automation platform for AI experts and consultants ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://useplumb.com/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠The Agent Readiness Audit from Superintelligent - Go to ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://besuper.ai/ ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠to request your company's agent readiness score.The AI Daily Brief helps you understand the most important news and discussions in AI. Subscribe to the podcast version of The AI Daily Brief wherever you listen: https://pod.link/1680633614Subscribe to the newsletter: https://aidailybrief.beehiiv.com/Interested in sponsoring the show? [e-mail]

Today on the AI Daily Brief, why workers do not trust their employers when it comes to AI. Before that, in the headlines, are we actually going to finally see an Apple AI acquisition? The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. Hello, friends, quick announcements before we dive in. First of all, thank you to today's sponsors, blitzie, super intelligent, and high touch. And to get an ad-free version of the show, go to patreon.com slash AI Daily Brief. Welcome back to the AID, The Brief Headlines Edition, all the daily AI news you need in around five minutes. And boy, we have a lot today. This is our last normal weekly episode of the summer.

Obviously in the U.S. on Monday, we have Labor Day, and that really kicks off the fall and back to school, back to work, all of the new energy that that brings. But this has been anything but a quiet summer when it comes to AI. And so appropriately, we are bursting at the seams. And we start today with the latest reports that Apple might finally be getting serious about an acquisition to solve how, woefully behind they are in AI. The information reports that acquisitions have become a serious topic of conversation among the C-suite, with senior VP of Services Eddie Q championing the idea. Apparently, rather than just licensing with Google to create AI Siri, Eddie wants to go much further. Sources say that, as we've heard in the past, the targets include mistral AI and perplexity,

and those conversations have been far more than just theoretical. Now, I will remind you that the reason that we talk about this is that historically Apple has not used acquisitions as a solution to get ahead in a new industry. In fact, to date, their largest acquisition is still beats headphones in 2014, which was just a $3 billion deal. In other words, like a quarter of what meta paid for less than half of scale recently. Of course, money isn't a big issue. Apple has about $75 billion in cash on hand,

and so has a credible path to buy just about any AI company besides OpenAI, Anthropic, or XAI. Based on the reporting, it seems that Mistral has been seriously considered at the $10 billion valuation they're currently seeking an venture, round, the information sources said that Eddie Q has been asking people inside and outside of Apple what they think of a mistral deal. However, he has reportedly been dissuaded because mistral is, quote, not among the top AI model makers. I tend to agree with that. I think at this point, a mistral deal would not be enough. I think markets would in fact reject it. I continue to think

that if they really want to play, Anthropic is where they would need to set their sights. However, at this point, I think that Anthropics' growth just might be too much for them to find a workable deal. What's more, while I don't think Anthropic was ever particularly keen to sell, I think now they have some serious wind in their sales and are even less likely to than before. Still, it is very much the case that every day that Apple waits, their options get more constricted, and the amount that they're going to have to pay does nothing but increase. As Beth Jaisos puts it, they got to stop talking and start writing checks. Speaking of big lumbering tech giants updating their strategy, Microsoft has taken a big step towards AI independence, unveiling a set of new in-house models. The two

models are MAI Voice 1, which is a text-to-speech model, and MAI-1 preview, which is a non-reasoning LLM. MAI Voice 1 is the focus of the press release and seems state-of-the-art-art-ish. We don't have any benchmarks so far, but the samples are reasonably impressive. Microsoft says the model can generate up to a minute of audio in under a second running on a single GPU, making it, quote, one of the most efficient speech systems available today. MA1 preview, meanwhile, is just a standard chatbot-LLM, really not much to say about it. The model is currently being benchmarked on LM Arena, and about the

The best you can say is that it's putting up a decent performance. It's currently ranked at number 13, which is below GPT-4-1 and GROC 3, but ahead of Gemini Flash 01 and Claude Sonnet 4 thinking, so essentially middle of the pack. Still realistically, the notable part is that it exists at all with Microsoft writing. This represents MAI's first foundation model trained end-to-end and offers a glimpse of future offerings inside copilot. Overall, Microsoft said, we're actively spinning the flywheel to deliver improved models. We'll have much more to share in the coming months. Frankly, I think Microsoft is playing a dangerous game here. I understand why in the wake of the absolute chaos of a couple years ago's board meeting and firing and rehiring of Sam Altman,

they feel the need to get more independence from OpenAI. But for their core business users, there is already a gap between the chat GPT that they're using with their Gmail accounts at home after hours and what's available to them as enterprise buyers using copilot. And unless they can seriously move quickly and get these models up to snuff relative to OpenAI's offering, and Google's offerings, et cetera, I think they are going to have a very difficult time. Speaking of Anthropic, a bunch of stories from them. First, they've released a new agent that lives in a Chrome plugin. Called Clod for Chrome, the product is a browser using agent that does pretty much all the things you would expect. In their announcement blog post, the company writes,

within Anthropic, we've seen appreciable improvements using early versions of Claude for Chrome to manage calendars, schedule meetings, draft email responses, handle routine expense reports, and test new website features. The company is treating this as a pilot test, releasing the agent to a selection of 1,000 Claude Mac subscribers. Anthropic does note that they're still very concerned with inherent security issues with web-use agents. Discussing the risk, for example, of prompt injection attacks. This is something, by the way, that has started to manifest in other web browsing agents

that were released earlier. Earlier this month, Brave published a security notice describing a functional attack vector, for example, in Perplexity's Comet Browser. Still, ultimately, Anthropic says that they view agentic web browsing as inevitable, and so they are attempting to strike a balance between safety and keeping up with the race. Maybe even bigger news from Anthropic is that the company has settled a copyright class action lawsuit with a group of authors. In a legal filing on Tuesday, Anthropic said that they had negotiated a proposed class settlement. We don't know the terms, but best guessing is that affected authors will see some sort of payout,

and Anthropic, for their part, gets to avoid going to a potentially costly trial. This lawsuit was filed last year and claimed that Anthropic used millions of books to build out their training data. In June, the judge ruled that Anthropics' use of copyrighted books in training data was fair use, which was a hugely impactful first-of-its-kind ruling. At the same time, the judge also found that Anthropic had pirated millions of books from the internet rather than purchasing them. Anthropic was set to go to trial over the issue in December, which had them facing fines that could have reached into the hundreds of billions of dollars

due to sheer volume. The settlement is expected to be finalized next week, so we will soon learn of the terms, but Justin Nelson, an attorney for the author, said, this historic settlement will benefit all class members. Now, it is worth noting that when it comes to broader implications, settlements can't be used as precedents in other cases, meaning that the direct impact is fairly minimal. At the same time, in practice, one settlement begets another, and so there could be some impacts there. IP scholars V Rosen writes, big news and a win for authors, I think. It doesn't

resolve the legal issues, of course, but I suspect we're going to end up with a licensing-oriented detente rather than a clear legal rule. Last Anthropic story, for the first time the company will be using user data to train their next generation of models. Coming after a big, change to their terms of service, the new terms will apply to free pro and max plans, but not the variety of enterprise and education plans or API access. Users can also opt out if they don't want Claude trading on their data. Anthropics said, by participating, you'll help us improve model safety, making our systems for detecting harmful content more accurate and less likely to flag harmless conversations. You'll also help future Claude models improve its skills like coding, analysis,

and reasoning, ultimately leading to better models for all users. Now, the Tech Press ran stories about how naive users might not read the fine print and accidentally opt into data sharing. I think, however, a more interesting question is what it means that Anthropic is shifting their policy at this stage in the game. OpenAI Insider Rune recently brought up the point that, quote, all model companies were pre-training on the same internet. Of course, Grock has access to Twitter dataset and Gemini can pre-train on YouTube and so on, but mostly it's the same internet. On the other hand, reinforcement learning environments will be whatever the lab chooses to prioritize, so you should expect more speciation. Indeed, Princeton professor

Arvin Narayanan writes, in an essay a year ago, we observed that more capable models don't necessarily mean more useful products and that real world usefulness will require training on user data. In retrospect, I'm surprised Anthropic held out this long. Speaking of this long, that is where we are going to wrap today's headlines. With that, we will move over to the main episode. This episode is brought to you by Blitzy, the Enterprise Autonomous Software Development Platform with Infinite Code Context. Blitzy uses thousands of specialized AI agents that think for

hours to understand enterprise-scale code bases with millions of lines of code. Enterprise engineering leaders start every development sprint with the Blitzy platform, bringing in their development requirements. The Blitzy platform provides a plan, then generates and pre-compiles code for each task. Blitzy delivers 80% plus of the development work autonomously while providing a guide for the final 20% of human development work required to complete the sprint. Public companies are achieving a 5x engineering velocity increase when incorporating Blitzie as their pre-IDE development tool, pairing it with their coding co-pilot of choice to bring an AI-Native STLC into their org.

Blitzy is providing a limited time, 30-day free proof of concept for qualifying enterprises. The team will provide a 5x velocity increase on a real development project in your org. Visit blitzie.com and press book demo to learn how Blitzie transforms your STLC from AI assisted to AI Native. That's BLITZY.com. If you are a regular listener, you will have heard about superintelligence agent ratingness audits at this point.

But I wanted to tell you today about the full suite of agent readiness products that go beyond just the initial readiness report. Over the last six months, Super Intelligence has built out an entire agent planning suite. We help you move from discovery to planning to implementation. After you've completed your agent readiness audits, we help you double-click on your most important use cases with what we call our use case planning reports. These reports are going to help you understand what sort of technical preparation you need to do to be ready for a use case, what challenges you might face in implementation, and whether you should be thinking about building, buying, partnering, or some combination.

After that, you can even get a spec document in what we call our technical blueprint that gives either your developers or the developers of the partner you work with what they need to build exactly the agent that you're looking for. If you want to learn more about superintelligence agent planning suite, we've built a custom GPT to answer your questions. Just go to bit.ly slash super super agent.

that's BIT.L.L.S. Super, super agent, all one word. And if you have any questions, the agent can even help you book an appointment with our team. Marketers spend endless hours building segments and journeys for campaigns. High Touch just announced a new round of funding to change that with their latest product, AI, decisioning. Instead of manually deciding which message goes to which customer, AI decisioning deploys agents that act like a personal marketing specialist for every single user. These agents learn from your own data warehouse, work across your existing tools like Salesforce Marketing Cloud, Brays, and Iterable, and continuously optimized for the outcomes you care about, like maximizing lifetime value. It's fully transparent, enterprise secure, and already in use by teams at brands like PetSmart, whoop, Weight Watchers, and Funrise.

So if you're ready to move past static campaigns and into the age of AI-driven marketing, check out High Touch's AI Decisioning at hightouch.com. That's hightouch.com. Welcome back to the AI Daily Brief. Today, we are talking about a really interesting and important phenomenon, some findings from a recent survey that suggests that when it comes to AI, employees do not trust their workplaces right now. However, before we get into that, I want to contextualize this in a broader conversation around the state of AI and jobs right now. A big theme for the last couple of months has been whether or not

and the extent to which we're starting to see AI show up in job displacement, particularly around young workers. You might remember back earlier in the year when Anthropic CEO Dario Amadee warned that AI could end up eliminating half of entry-level white-collar jobs, and giving his standing in the industry, obviously there was a lot of discussion around that. So far, it's been a lot of tea leave reading and theorizing, and so I was interested to see this recent paper come out from Stanford University's Human-Centered Artificial Intelligence Lab or H-AI. The paper took a targeted look at how early career workers between the ages of 22 and 25 are being impacted by AI. Now, whether or not it is AI's fault. By some metrics, this has been the worst year to leave college since

2015. NBC recently published a piece called A Black Hole, New Graduates Discover a Dismal Job Market. They write NBC News asked people who recently finished technical school, college, or graduate school, how their job application process was going, and in more than 100 responses, the graduates described months of spent searching for a job, hundreds of applications, and zero responses from employers. Now, obviously, a news outlet asking 100 people does not a comprehensive data set make, And yet there's definitely a broader sentiment that this is happening in some ways. If you go on various community spaces online, you'll see people complaining about similar phenomenon. What's more, as NBC points out, the unemployment rate among recent grads has been increasing this year

and is up to about 5.3% as compared to around 4% for the labor force as a whole. By the numbers, then, it's one of the toughest job markets for new grads since 2015. Of course, what's difficult is to understand how much this is about AI specifically versus confounding factors like the reversal of hot tech hiring in 2021 or just general malaise in the private sector. Now, the Stanford paper does try to control for other factors to come up with a definitive answer on whether AI is already impacting the labor market and ultimately came to the conclusion that yes, AI is having an impact here. The paper highlighted several key findings. Their main one, and the one that you've seen the most statistics probably shared on places like

X, is that early career workers entering fields that are highly exposed to AI disruption experienced a 13% relative decline in employment even after the data was controlled for firm-level shocks. The researchers gave the examples of customer service representatives and software developers as some of the professions most exposed. In contrast, employment for workers in less-exposed fields saw employment remain stable or even grow. The same was true for more experienced workers in occupations that are highly exposed to AI. You can see this chart that was shared by Mike Byrd from The Economist that broke different groups into their exposure quintiles and seemed to find this pattern of higher AI exposure, leading to lower job growth or actual

job deceleration. The researchers also found that adjustments in the labor force are more likely to occur right now through headcount reduction rather than reduced compensation, and reductions in the labor force were more likely in fields where AI was being used as automation rather than augmentation. In other words, and intuitively, voice agents taking over for call center workers is a challenge, but AI note takers don't seem to pose much risk. Finally, the researchers said that these effects were still visible after controlling for several different factors and remaining consistent across alternate sample constructions. They said, we find that the AI exposure taxonomy did not meaningfully predict employment outcomes for young workers further back in time before the

widespread use of LLMs. The patterns we observe in the data appear most acutely starting in late 2022 around the time of rapid proliferation of generative AI tools. In other words, these patterns are new and at least correlated with the rise of AI. Now for all of this, the paper is still careful to caveat that the results could be due to factors other than AI. Ultimately, they reduce their claim to state, our results are consistent with the hypothesis that generative AI has begun to affect entry-level employment. Still looking into the data, it really does seem like there are a lot of confounding factors here. Marketing and sales managers is one of the occupations that researchers suggested was most impacted by AI. However, the biggest drop in junior hiring came in

the last quarter of 2022. So then if you extend the logic of the paper, you'd have to be arguing that effectively, chat GPT gets released in November and firms immediately stop hiring junior sales and marketing positions? If you ever use GPT3, you'll realize just how absurd that notion is. Hiring in that field also seemed to recover strongly in the second quarter of 2024 when the economy got out of its slump. The Contra example is also a little flawed, with the researchers using nursing aids as the key example of an occupation where young hires are growing. It's probably true that nurses haven't been disrupted by AI, but that's also a field seeing a huge demand spike due to the aging occupation. This is of course not to argue that AI is having no effect, but to try to

claim that because we are starting to see these patterns in young employment, the only or even main causal factor still might not just be AI. Ultimately, we are still very much in the vibes and sentiment stage of this, not the hard data stage. And I do think that as opposed to some recent so-called studies from top universities that make big bold pronouncements about AI. In this case, the Stanford group is just trying to get a little bit of color to start to try to see some signal through what is ultimately very noisy and immature data. I think it's also important that any time we talk about AI job loss, we also talk about jobs that AI will create. And this is not just me blithely being an industry advocate, although I've certainly been accused of worse in my life.

It's that anytime we talk about job displacement from new technology, it's so much easier for us to understand what things that we do now that new technology will be able to do instead of us, as opposed to what new things we don't even consider doing now, that technology will create the opportunity for. That is definitionally in the future and harder, if not impossible, to know. And on that front, a new report from staffing firm, Birchworks, found that AI-specific roles are absolutely on fire. The report found that base salaries for non-managerial workers in AI with zero to three years of experience grew by 12% over the past year. BurchWorks' and analyze the compensation of thousands of AI and data science candidates to come up with these numbers.

The report also found that people with AI experience are being promoted to management roles roughly twice as fast as their counterparts in other tech fields. Enil K. Gupta, a professor at the University of Maryland's business school, said, there is a significant salary difference between a machine learning engineer job and a software engineer job. The WSJ added some anecdata from Database CEO Ali Goatsy to confirm. He said that he plans to triple graduate hires this year in part for their AI skills. He said, they're going to come in, they're going to be all AI native. We can't, for the life of us, get the more senior people to adopt it.

Now, interestingly, at the Enterprise AI event that I was speaking at this week in Vegas, a lot of the conversation in some of the breakout sessions was about this phenomenon. I heard a lot of the people who were here in positions of power, CIOs and the like, talk about how they were really excited about young people who were more AI native coming in and helping bring that skill set and that mindset to the organization. I think there is an awareness that many firms are going to treat junior workers like they are not important anymore because of AI, and some other opportunistic firms are going to swoop in, treat them really well, and potentially get the benefits. All of which brings us to this interesting conversation around the relationship between workers and their employers when it comes to AI.

If you're not familiar with Kyla Scanlan, she is a economics-focused content creator who's got a big following on TikTok and on other social media and does a great job of explaining and breaking down important economic and market concepts in a way that non-professionals can understand. A couple of days ago on her blog, she published a survey that she had run called AI That Works for Workers. She writes, AI is at least four different transformations happening simultaneously, in how we work, how we invest, how we power our world, and how nations compete.

We're trying to integrate an unpredictable technology into systems that weren't designed for it. What's happening in this messy middle? I wanted to know how people felt about AI. What do workers think about working alongside these new systems? So, I asked them directly. So basically, this is a self-conducted survey. It is not a Stanford survey. It's not even an MIT survey.

But Kyla has a big following, and she got 1,200 people to respond to the survey within 24 hours. As she puts it, the entire goal here is exploratory research. This is more meant to be a preliminary exploration of how people are thinking during this time of immense change than to produce declarative conclusions. The survey that Kyla shared had 11 questions. How familiar are you with AI tools in your research? industry? Have you used AI at work in the past six months? What's the biggest benefit you hope AI could bring to your work? What are your biggest concerns about AI at work? What role should workers have in decisions about how AI is implemented? How much do you trust your employer or organization to use AI

in ways that benefit workers? As your employer provided training on how to use AI tools? If you could enact one policy to ensure AI benefits and complements workers, what would it be? And then finally, what industry are you in? Are you a member of a union? And any additional thoughts. She writes, their responses largely showed a workforce that's not totally enthusiastic nor completely resistant, but they are actively thinking through what it means to work alongside intelligence that isn't quite human but isn't really predictable either. Now, there is a lot of interesting stuff in this survey, which, by the way, you can go check out at kaila.substack.com. That's kyla.com. But by and large, the thing that I found interesting was the utter lack of trust that employees have with their

employers when it comes to this technology. Now, this is not the first time that we've seen there be a gap between employees and executives when it comes to AI policy. A number of times on this show I've shared this survey from writer that was conducted last December that surveyed 800 employees and 800 C-suite executives and saw, frankly, a huge gap between how they were thinking about AI and their own companies. When asked whether they thought that their company had been successful in adopting AI over the previous 12 months, 75% of executives said that they had versus only 45% of employees. In fact, when it came to the simple question of just asking whether their company had an AI strategy,

whereas 89% of executives said, yes, of course we do. Only 57% of employees said the same. Now, there are all sorts of interesting places that this lack of trust is showing up and suggesting that there are real problems here. That same survey found that 41% of Gen Z employees are sabotaging their company's generative AI strategy, and that of the people who are, a third of them are worried about AI diminishing their value or their creativity. point being, there appear to be big chunks of people inside companies who are very much not on board with AI strategy. And it's understandable why. If you think that when a company is pushing an agent on you, they're basically asking you to just train your replacement, you're not going to be all that enthusiastic about it. Which is not to say that there aren't things that people want out of AI.

When Kyla asked what workers hope AI will do for them, 29% said that they wanted to reduce repetitive work and 27% said that they wanted to increase efficiency. and yet they worry that it could shrink their opportunities and that the early state could actually impact negatively the quality of their work. 18% said, for example, that they were concerned about fewer career opportunities. Still, as I said to me, the most interesting numbers were around the trust or lack thereof between employees and their companies when it came to AI. The chart that you're looking at shows a breakdown by industry of how workers rated how much they trust their employer to use AI in ways that benefit them, the workers.

On the more trustful end of the spectrum, you had the healthcare industry. 10% of workers in that industry said that they completely trust their employer, and 61% said that they had some trust. 29% said that they had no trust. Now, I'll pause here to note that in the most trusting industry, there were still about 30% of people just under a third who had no trust in their employer to use AI in ways that benefit them. On the other end of the spectrum, we have the real estate industry, where 0% completely trusted

their employer to use AI in ways that benefit them, and only 44% had some trust, as opposed to 56% who had no trust. The vast majority of industries here had over a third of people with no trust in their employer to use AI in ways that are beneficial to them, as opposed to just for the company, and just under half of them have over 40% saying that they have no trust. So if you are an executive in thinking about 2026 and how you get AI working inside your organization, it may be useful to consider what your employees currently think about how your organization is going about things. Now, part of the reason that people might not be trustful is that it doesn't necessarily really seem like they're being supported when it comes to AI. The vast majority of workers,

for example, in this study, had never received any training on AI. The two industries that had the greatest percentage of people responding that they had received some training was the technology industry itself and consulting. And in those cases, it was still only around 60-40 of people who had been trained to people who hadn't. In consulting, it was 5842. That means that in consulting, an industry that is so poised for disruption or at least radical change from AI, you still have 42% of workers who had never received any training from their employees on AI tools.

And by the way, they want it when asked for their thoughts on what their employer could do to increase their comfort with AI and their positive relationships with AI, training and upskilling funds was the top answer. To me, it seems inevitable that the more that our focus turns to agents, and especially highly complex agents or systems of agents that can do bigger and bigger chunks of work, the more that you're going to see human employees push back and demand to have a seat at the table. Now, the opportunity, of course, is that for forward-thinking organizations, while other organizations wait too long and struggle with things like employee sabotage,

organizations that actually get their people on board, that paint a vision for how AI and agents are going to work at their company and what the implications are for their existing employees, those organizations are just going to so dramatically outperform those that don't do that. Of course, we are still living in a dearth of real data, but as long as the vibes keep rolling in, we will keep keeping track of them. For now, that is going to do it for today's AI Daily Brief. Appreciate you listening, as always, and until next time, peace.