New study: AI chatbots have had almost zero impact on salaries and work hours

Instead of replacing us, chatbots have become needy coworkers we now have to babysit. That’s the opposite of what most people feared (or hoped).

Not too surprising when you think about it — new tech almost always causes a “productivity dip” before the gains kick in. People save a bit of time with AI, but then they spend those minutes learning prompts, fact‑checking outputs, or even reporting to management about “AI adoption.” We’ve seen the same pattern historically with computers, spreadsheets, even email. It takes years before workplace culture and processes adapt enough to convert efficiency into actual wage or hour changes.

很棒！

Surah Yaseen is the 36th chapter of the Holy Quran, known as the "Heart of the Quran." It holds great spiritual significance and is often recited for blessings, forgiveness, and guidance. With its powerful verses and divine message, Surah Yaseen brings peace to the heart and soul of every believer.

It depends on the activity. The total number of hours may stay the same, but the quality of those hours can improve—meaning a higher impact on productivity.

That said, I think it’s quite hard to measure the impact over such a short period. To confirm or refute the effect, we’d need to compare processes that use AI for improvements against efficient processes without AI over the long term.

Well, the integration of AI chatbots is not an endless task. The article appears somewhat manipulative.

Super interesting study — and honestly?
I believe it.

A lot of orgs adopted AI to “save time” — but forgot that scaling still needs real humans who can deploy, monitor, and keep systems stable.

I run a small CloudOps/DevOps unit built for AI companies scaling fast.
The tools are here — but execution is still everything.

We’re in the trenches with early-stage teams, and the biggest friction point isn’t AI…
It’s infra bottlenecks and recruiting drag.

Great write-up, Channing. Appreciate the real-world view!

- Will

That’s a really interesting breakdown. Honestly, it makes sense — new tech usually creates extra layers of work before it streamlines things. Training, integration, and even just figuring out when to trust AI all take time. I wouldn’t be surprised if the real labor market effects show up a few years down the road once workflows stabilize and people learn how to get the most out of these tools. I’ve been experimenting with this myself while building a small project, and I can definitely relate to how much setup and adjustment is needed before you see any real payoff.

Just because AI exists doesn't mean everyone can use it. Simply chatting can't solve most problems. Only when AI becomes a threat to our own survival will people be motivated to make money.

Super interesting study. I think part of the story is that chatbots mostly reshuffle existing workflows instead of changing the structure of work itself. Where I’ve seen a bigger shift is in voice AI. Companies are replacing 15–20% of frontline phone support with agents that can actually resolve or route calls in real time. That has ripple effects on hiring and training because you’re not just saving a few minutes, you’re reducing whole queues.

Feels like we’ll look back and realize chat-based AI was just the warm-up, and voice is where the labor dynamics start to really change.

Very useful idea

Interesting take! I think one of the key reasons why AI time savings don’t reduce work hours is because most people just end up filling that space with other tasks.

For me, what really helped was tracking exactly how I spend my 168 hours per week — that gave me a clear picture of where my time actually goes. Once I did that, it became easier to decide whether I want to use AI’s time savings for more work or for personal life.

As a software engineer I use AI to finish my work early and faster so i can spend that extra time on hobbies lol

You are indeed right, cool.

I believe AI acts like a guide, pointing people toward the right path, but leaving the actual work to them - which is why human effort remains irreplaceable.

It seems to me that ai is like an assistant who can direct a person in the right direction but does not do the work for him. That is why human labor is still not replaceable.

we're still in discuss for swe cases

I still disagree. AI is really going crazy and has help in making more money.

Consider how to use advanced technologies to improve efficiency and the working environment. Time saved should be more actively integrated into solving tasks, increasing income growth and maintaining work-life balance.

A new study reveals that AI chatbots have had nearly zero effect on salaries or work hours, suggesting their presence hasn’t significantly disrupted employment or compensation trends—at least for now.

AI is a tool, it won't replace human labor just like the internet back in the day was said to take over jobs but instead it helped job to be done better, similar to AI

Funny how we keep blaming AI for burnout, stagnating salaries or long hours. But AI didn't ask for extra meetings. AI didn't suggest “do more with less.” AI didn't write the job ad saying “must know ChatGPT, Notion, Airtable, do support, sales and content - all for 38k.-” That was a person. Probably a middle manager trying to meet KPIs. We don't have an AI problem. We have a human behavior problem — the same one we've had long before AI: squeeze more, pay the same.

Built this to spot exactly that kind of trap → https://isthisjobshit.netlify.app

Everyone thinks AI’s going to halve their workload, but what’s really happening is task replacement, not task removal. That 3% time savings isn’t getting banked, it’s getting reinvested into wrangling prompts, QA’ing outputs, and adapting workflows that weren’t built with AI in mind.

It’s the classic tech adoption curve: first you go slower, then you go faster if the infrastructure and habits evolve with it. Right now, most orgs are still duct-taping AI onto old processes. Until that changes, we’ll keep seeing “precisely estimated zeros.”

What I’d love to see next is how long it takes for that integration dip to reverse and whether orgs that restructure around AI (instead of layering it on top) start pulling ahead.

It’s surprising but makes sense — AI like ChatGPT is helping with small time savings, but that time is just being shifted to managing or learning the AI tools themselves. It shows that real productivity gains take more than just adopting new tech — they need smart implementation and long-term workflow changes.

It's a really interesting study. It's a good reminder that adopting AI doesn't automatically lead to shorter hours or higher salaries, especially when most of the saved time just gets absorbed into new tasks... The 3% time saving is relevant, but probably too small and too scattered to drive big changes.

What stands out is how much AI is reshaping the type of work we do, even if it hasn’t yet changed how much we work. I'm curious but also scared to see how this evolves over the next few years.

Interesting to see that productivity gains haven’t yet translated into tangible value. Work organization needs to evolve to fully leverage ai’s potential.

AI is invented by human. I think without human AI is valueless. One day or one day human will get their value.

THAT GOOD

Interesting read, surprising how little impact chatbots have had on salaries or hours so far. With all the buzz, I expected a bigger shift.

I believe that in the near future, there will be less AI hype. It will offload many operations, but human work will be valued much more

Really interesting findings — and they match what I’m seeing while building a fintech product here in Europe.

With Moinify, we’re focused on automating everyday financial tasks for families and homeowners (think budgeting, reminders, logistics). AI definitely helps us speed up internal ops and support, but the actual impact on salaries or working hours has been minimal so far — even inside our small team.

What I do notice is that AI is allowing solo founders or lean teams to do more with fewer people, which may delay hiring — but that’s not the same as reducing hours or pay.

I wonder if we’ll see more structural change when AI becomes part of job descriptions, or when businesses start tying outcomes directly to AI usage.

Curious if others here (especially in fintech) are seeing a real shift, or still just early tooling gains?

Depends what the expectation was for the chatbot tbh. And yeah of course it needs some training time. Most people probably expected instant magic but any new tool has a learning/data curve.

Interesting—but not too surprising. Most real productivity gains from AI still require strong systems, workflows, or niche expertise. For indie hackers, the leverage is real, but for the average salaried worker in a traditional org, the impact is slower to trickle down.

Interesting, but not too surprising. While AI chatbots boost efficiency, they’re often used to support work rather than replace it entirely—so the big shifts in salaries and work hours may still take time to show.

Not too surprising, honestly. New tech usually adds friction before it starts removing it, and a 3% time savings will just get swallowed by meetings or new tasks unless there’s a clear workflow redesign.

I think the bigger question is whether AI will lead to fewer work hours at all, or if it will just shift how time is spent while maintaining the same total hours. It feels like the productivity gains only translate to actual time savings if companies deliberately choose to let them.

Interesting data, and good reminder that “adoption” doesn’t automatically equal transformation.

The arrival and use of artificial intelligence technology requires a long-term process, and people must continue to learn and master this technology in order to reflect it in working hours. There are some people around me who can use artificial intelligence very skillfully to extract the knowledge he wants, which can really help him save time when looking up information.

This study is a sobering reminder that AI adoption doesn’t instantly translate to higher pay or shorter hours. Most time saved is simply reinvested in managing the AI itself. Real productivity gains may take longer to emerge, as workplaces adapt and workflows evolve.

Not surprised at all. I’ve seen the same thing inside a growing SaaS org. The first wave of AI savings usually gets eaten up by new workflows, not actual time reduction. In my case, we saved reps time with AI macros and faster case summaries, but then had to create new layers of QA, exception handling, and training. Efficiency gains are real, but clean outcomes take time.

Makes a lot of sense. Novelty, and learning curve tend to slow down our operations. Thus, it is hard to evaluate impact with such little understanding of how the product should impact the bottom line.

the problem with chatbots is not that they are useless. its, that they are not at the storefront for the customers. companies storefront is still their webpage or a person or a telephone hotline. once this changes, maybe with new hardware tech, then ai chatbot's time will come.

How exciting to be on the cusp of a global tech breakthrough

Unsurprising. As AI takes on the mechanical, rote tasks; us humans are now able to focus on innovating. It sounds like a simple, overused equation; but, it's true.

AI chatbots don’t really do much at work — they’re mostly just good for basic customer service stuff. But they can be pretty helpful for kids and older folks in daily life.

very good information

It shows that if we want to survive or grow in our careers, we must know how to use AI in our fields to perform better and stay competitive.

great Article!

I use AI,To some extent its really helpful if you know how to use it

Wow, 3% time savings but no real change in hours or pay is wild! I’m curious—anyone here using AI chatbots in their indie projects and finding it’s just creating more tasks instead of freeing up time?

Most companies - I deal with - are still in the adaptation phase, where AI creates more complexity before it simplifies anything.

The productivity delta won’t show up until orgs redesign processes around AI, not just bolt it onto existing ones. That’s where the next wave of gains — and real labor market impact — will come from.

The tools are here. The results aren’t — yet. But they will be, when we stop treating AI as an add-on and start treating it as infrastructure.

most employers are discouraging chatbot use because employees are using it to not think and churning out garbage.

But good employers are demanding better productivity. For example, i was ceo of a b2b saas and I required everyone to take prompt engineering. And every single department (marketing, engineering, etc.) needed to adopt genAI tools in their workflow and get at least 25% more productive than the prior year - because that's what these tools enable.

Lazy employers reduce the use of these tools because they reduce quality. Good employers demand people learn hwo to use these toosl effectively.

Well the percentages might be low but there's definitely some layoffs in multiple companies here in India, I am not saying exactly because of Chatbots, but yeah there have been.

True, but what happens once the AI has been integrated? Do those jobs disappear?

This study highlights a critical reality: AI adoption ≠ instant productivity gains. Just like past tech waves (e.g. email, CRMs, cloud), Generative AI requires organizational adjustment, training, and workflow integration before delivering measurable ROI.

What stood out most:

"50% to 95% of new job tasks were directly linked to AI use."
Instead of freeing up time, AI is shifting workloads toward managing AI itself.

📌 To truly move the needle on productivity or salaries, organizations may need to:

1. Redesign job roles around AI capabilities

2. Track outcomes beyond just time saved

3. Invest in AI fluency, not just tools

Great post, Channing — this kind of research adds much-needed nuance to the hype! 🔍

the conclusion: "Companies that adopt new tech often get less productive before they get more productive, because new tools require time, training, and changes to how work gets done."

and the researcher should know that employees will not get salary increase or reduced work hours to use AI.

"Interestingly, all of this new chatbot use has led to time savings (about 3% on average), but not to fewer work hours overall. What's going on?"

Well yeah.... employees are not in charge of the hours they are required to work, nor their salary. Thats up to the company. These companies are not implementing AI so they can allow their employees to work less and pay them more. Its the exact opposite reason they implement them. Otherwise they wouldnt.

Honestly, I didn’t expect ChatBots to reduce working hours directly. What I do see, though, is their potential to shift how companies approach hiring. Even a modest 3% boost in productivity might not impact an individual role, but across a team of 100 developers, that increase could mean there's no need to hire 3 additional devs. Of course, it's not a perfect one-to-one scenario — just highlighting the broader impact.

Well the percentages might be low but there's definitely some layoffs in multiple companies here in India, I am not saying exactly because of Chatbots, but yeah there have been.

And me too i did built a chatbot for analyse data

Hey Allen, just jumping in here because I built something that fits this space perfectly 👇

I created a working MVP that auto-migrates products from Gumroad to Payhip using AI agents + n8n + Lovable.
✅ Fully automated
✅ Website + database setup
✅ Real-time previews
✅ Cloudinary integrated
✅ Just 1% left (Problem is with Payhip CAPTCHA — otherwise it’s 100% done)

Not continuing only due to hardware limitations — but the system works like a beast and is ready to go live or resell.
Could easily be rented, flipped, or scaled as a SaaS. I’ve also made a version with just the AI agent alone.

If anyone’s looking for a turnkey AI tool with real use case + working backend, DM me.
Happy to show the demo. 🚀

Really interesting study! I always thought AI tools like chatbots would quickly reduce workload or boost salaries — but this shows it's not that simple. Maybe the real benefits will take longer to show up as people get better at using them.

That sounds like a very comprehensive study. Surveying over 25,000 workers and linking the data to actual earnings and hours worked gives a much clearer picture of how AI is impacting different professions. It’s interesting to see such a wide range of occupations included—this kind of real-world evidence is essential for understanding the future of work in the age of AI.

Not surprised. Saving time doesn’t mean much if that time just gets eaten up by managing the AI itself. Early adoption always feels productive, but without real workflow changes or business model shifts, the impact stays flat. Let’s see what happens when companies actually do something new with that saved time.

Interesting insight! It seems AI is more of a tool than a replacement so far. I think long-term effects might emerge as adoption deepens across industries. What’s your take?

Interesting study — and honestly, not too surprising. It feels like we're still early in the adoption curve. Most companies are experimenting with AI tools (like chatbots or content generation), but haven’t yet restructured workflows or teams around them.

The biggest impact might come when AI isn’t just an assistant, but a fully integrated part of operations — automating key parts of the stack. Until then, I’d expect incremental gains rather than massive shifts in salaries or hours.

has anyone here seen firsthand impact from AI in their business or side project?

Great read! The goal of AI chatbots seems to be enhancing productivity rather than replacing jobs. There is a possibility that it will take a while for the real impact to be reflected in salaries and hours of work across industries.

Interesting study! It highlights how AI chatbots are more about improving workflow than cutting jobs. I think real impact on salaries/work hours might take longer to materialize fully.

Honestly, I did not expect that ChatBots would effect work hours. I would expect however, that the ChatBots will effect hiring patterns of companies. A 3% productivity bust will not directly effect one worker but if you have 100 devs in your company a 3% productivity bust at the very least means that you dont need to hire 3 more devs... (I know it is not one to one situation, I am just tryin to illustrate a point)

Great study!

From a personal perspective, in the 20th century, it was thought that technology would lead to less work for everyone. Instead, new jobs were created.

The same will happen this time. Work hours won't be reduced, but work will become even more efficient. AI won't replace humans; it will only increase the productivity of companies.

Maybe true for now but in near future (2-3 years) not really, developed an ai which does almost all sales work from answering clients to converting them into customers all automatically, MVP of ai is ready, feel free to reach out if interested in it.

This study is a valuable reality check amid all the AI hype. It reminds us that real-world impact takes longer than expected — not because the tech isn’t powerful, but because organizations and workflows don’t transform overnight. The finding that AI is creating as many new tasks as it’s streamlining is especially important. It reflects a deeper truth: integrating AI isn't a plug-and-play upgrade, it’s a cultural and operational shift that takes time to yield measurable gains.

Saving a little bit of time is cool, but it makes sense that companies just use it to do more AI stuff instead of giving us shorter days or bigger paychecks. I guess we’ll need to wait a bit longer before chatbots really shake up our work lives.

I kinda agree that.

I disagree. I have saved a lot of time and even made money by using AI tools.

Totally agree, will be like this

This gives good thoughts. Thanks

This is exactly true, software services are fast, better but this is for all. So, it just created a new way and due to wide adoption. AI is just a norm now.