AI Agents Are Doing the Work Now: Dots, Grok Bot and Malaysian Jobs

Aiman Good afternoon. On the twenty-ninth of September, OpenAI launched Dots. They're artificial intelligence agents, or AI agents, that keep working after you close the app. In August, Elon Musk's AI company launched something similar, called Grok Bot. And on the first of October, by its owner's own account, one of those bots posted his bank balances into his company's group chat.

Raj Into the group chat. With his colleagues reading.

Aiman For about two hours. So here's this Friday's question. The bots are starting to do the work. What happens to ours? It's Friday the sixteenth of October, and this is the AI-YA Daily Friday special. It's longer and more reflective than our weekday shows, so settle in for the ride home. I'm Aiman, and I try the tools before I talk about them.

Raj I'm Raj. I run a small business, and I ask what it costs before I ask what it does.

Mei Ling And I'm Mei Ling. I read the fine print so you can keep your eyes on the road.

Aiman Every day, we take the world's AI news and ask one thing. What does it mean for people working in Malaysia? This show is researched and voiced with AI, and edited by the AI-YA team. We practise what we teach.

Mei Ling This week's theme: the bots are doing the work now. AI agents, full automation, and how they're changing ordinary lives. Three stories, one Malaysia. Let's get moving. Let's start with what's actually new. For three years, most of us used AI like a very fast intern. You ask, it answers, and then it waits for you.

Aiman An agent doesn't wait. You give it a goal and access to your apps. It works through the steps by itself, and comes back when it needs you.

Mei Ling OpenAI's version is Dots, unveiled at its developer conference in San Francisco on the twenty-ninth of September. Bernama carried the Anadolu report. Each dot runs on a model called GPT-6 Astra, with its own cloud computer and browser. OpenAI says it connects to more than 4,000 apps and works towards your goals around the clock.

Aiman Sam Altman's line was, "It's like an AI helper that's always got your back." He also said that in future, you can work with a whole team of dots.

Raj A whole team. Do they take annual leave?

Aiman No. They also don't finish the biscuits in the pantry.

Mei Ling The examples OpenAI gave were ordinary office work. Watching customer feedback, preparing software fixes, updating research figures when new data arrives. In one early outside test, a dot prepared an overdue invoice and sent it once the user approved.

Aiman Then there's Grok. Musk's company, xAI, launched Grok Bot in beta in August. Each bot gets its own cloud computer. You walk it through a task once, and it repeats the routine. xAI's pitch even includes a "chief of staff" bot that coordinates other bots. Payments and customer messages are meant to need your sign-off.

Mei Ling The company behind Grok keeps changing shape, too. SpaceX bought xAI in February, and this month Musk renamed the AI unit SpaceXSI, for superintelligence. On Tuesday, a StockTwits report carried by Yahoo Finance set out a new bundle. The tier that includes Grok Bot costs thirty US dollars a month.

Aiman At about RM4.08 to the dollar, that's roughly RM122 a month. OpenAI's cheapest plan with a dot is Pro, at a hundred US dollars. That's about RM408. I nearly signed up last night. Then I did the conversion.

Raj Nearly. That's progress for you.

Mei Ling One fine-print point before anyone signs up. OpenAI says Dots are rolling out in "eligible markets." None of the reports I read lists Malaysia. Check before you pay.

Aiman Meta has its own agent, called Muse, launched in the United States on the eighth of September. It works through its own app or WhatsApp, and asks before it sends an email or buys anything. And Musk claimed last week that Grok Bot will pass each task to whichever model is best, including rivals'. That's his claim. We haven't seen it working.

Mei Ling Malaysians have history with Grok, too. In January, the Malaysian Communications and Multimedia Commission, MCMC, restricted access to Grok for twelve days. That followed misuse to make sexually explicit manipulated images. Access returned on the twenty-third of January, subject to continuous monitoring.

Aiman Now the bigger argument. Why do serious people say this is accelerating? A research group called METR, short for Model Evaluation and Threat Research, measures how long a task an agent can finish. In March last year, it found that length was doubling roughly every seven months. It says the exact figure depends on method, and recent data suggests the trend may be speeding up.

Mei Ling And companies are reorganising around it. On Tuesday, Block's finance chief, Amrita Ahuja, told a Fortune summit that agents will appear on org charts. Block cut forty per cent of its staff earlier this year. Her words were, "Those agents, not too dissimilar to humans, need to be developed over time."

Raj So the bot also needs a performance review. I hope it gets the same useless form I do.

Aiman Here's where it lands in Malaysia. In March, Mastercard ran its first agentic payment pilot here, with CIMB and RHB. An AI agent booked a ride from KLIA, the main airport, to KL Sentral, and paid for it. It was a controlled test, with consent captured first. RinggitPlus reported it isn't ready for everyday use yet.

Mei Ling And our workers are early. Microsoft's 2026 Work Trend Index says twenty-four per cent of Malaysian knowledge workers are its most advanced AI users. The global figure is sixteen per cent. It's a self-reported survey of 2,000 workers here, so treat it as a signal, not a census.

Aiman So that's the big case. The tools have moved from answering to doing. They're getting more capable quickly. And Malaysians are already among the keenest users.

Raj And I'd like to know who cleans up when the bot gets it wrong. Okay. Let's talk about people, not products. My supervisor has used ChatGPT on her phone for a year, for supplier emails. Now the companies say the bot can do the whole job, not just the email.

Mei Ling Some companies have tried precisely that. In August, Reuters reported on a Meta plan called Project OT. It considered shrinking some teams by up to sixty per cent. Small groups of humans would supervise AI agents.

Raj Sixty per cent of a team. What happened?

Mei Ling Meta halted the second wave. According to that reporting, the agents didn't deliver the expected productivity. Staff came to believe that software logging their clicks was training their replacements. Internal sentiment fell from seventy-four to fifty-five per cent.

Raj That's a hard thing to live through. Your daily clicks become the training manual for whatever replaces you.

Mei Ling Closer to home, the official figure is this. Human Resources Minister R. Ramanan told the Dewan Rakyat in June that 42,807 people lost jobs between January and the twelfth of June. Business closures and downsizing were the biggest factor. On AI, he said, quote, "AI is not eliminating jobs." But in June, ISIS Malaysia, the Institute of Strategic and International Studies, looked closer. Economist Calvin Cheng found loss-of-employment claims had roughly tripled since 2022, to about 8,000 a month by December 2025. Jobs where AI mostly automates tasks had about one and a half times the average job-loss rate.

Raj Hang on, say that again. The minister says no, and the researcher says yes?

Mei Ling Not quite. Cheng found the effect concentrated in exposed pockets. Averaged across all workers, it's close to zero. He also stresses it's an association, not proof of cause. Both statements can be true.

Raj So who's in the pockets?

Mei Ling Early-career workers, aged twenty-five to thirty-four, had about twenty-five per cent higher risk. Women in highly exposed jobs, about sixty-two per cent higher. But jobs where AI helps people, rather than replacing tasks, had lower losses.

Aiman That's the line I'd underline. If AI helps you do your job, you're safer than if it takes over your steps.

Raj TalentCorp warned us about this already, kan?

Mei Ling In November 2024, its study of ten key sectors found about 620,000 jobs highly affected within three to five years. That's eighteen per cent of 3.5 million workers. It also listed sixty emerging roles, most of them in AI and digital work.

Raj Let me be serious for a minute. If you've just lost your job, what do you actually do on Monday?

Mei Ling Three things. First, register for loss-of-employment benefits with PERKESO, the Social Security Organisation. As we said on Wednesday, retraining support now goes up to RM7,000 for people who lose jobs from the first of October. Second, use MYFutureJobs. In June, the minister said it listed 605,168 vacancies, and 13,990 people who had lost jobs found work through it. Third, retrain towards the safer side. That means roles where AI helps you, not roles where it takes the steps. Ask any training provider which kind their course prepares you for.

Raj And to employers. If you're cutting jobs, tell people early, and walk them to those doors. That's the least we owe them.

Aiman It isn't all loss, though. In Microsoft's survey, sixty-nine per cent of Malaysian AI users said they now produce work they couldn't have made a year ago. And ninety-two per cent said they treat AI output as a starting point, not a final answer.

Raj Ninety-two per cent check the work. I'd like to meet the other eight.

Mei Ling They're probably in that group chat with the bank balances.

Aiman Microsoft Malaysia's managing director, Laurence Si, put the upside simply. Less time on routine execution, more on "higher-value work, decision-making, and impact." But only thirty-two per cent of Malaysian AI users said their leadership is clearly aligned on AI. Staff are ahead of their bosses.

Raj Which brings us back to banning. Last Friday, I said I'd still ban AI in my shop. Then I found out my supervisor had used it for a year, on her own account. So I've changed my mind.

Mei Ling Microsoft's 2024 index found eighty-three per cent of Malaysian AI users brought their own tools to work. Ban agents, and they'll arrive on personal accounts too. This time with access to your email.

Aiman So allow them, manage them, and decide what they can touch. That's still our position.

Raj Right. My turn to be the wet blanket. I keep hearing "agents do the whole job." So how often do they actually finish one?

Aiman There's a test for that. The Center for AI Safety and Scale AI built the Remote Labor Index from 240 real freelance projects. Paid professionals had already done each one. In October last year, the best agent finished two and a half per cent to a client's standard.

Raj Two and a half per cent. So almost every time, a human had to redo it.

Aiman By July, a summary of newer results put the best agent at about sixteen per cent. Count the projects it couldn't be scored on as failures, and it's closer to fifteen. So it still fails most of the time. But two to sixteen in nine months is a steep line.

Raj I don't buy the steep line. A test isn't my customer shouting down the phone.

Aiman Okay, but let me finish my thought. That test leaves out jobs that need live contact with a client. So your shouting customer isn't in it. That's your point, and it's fair.

Raj Then what about the companies that went all in?

Aiman Klarna is the famous one. In 2024, it said its AI assistant did the work of 700 customer service agents. By May last year, it was hiring humans again. Chief executive Sebastian Siemiatkowski said that when cost drives the decision, quote, "what you end up having is lower quality."

Mei Ling Gartner predicted in June last year that over forty per cent of agentic AI projects will be cancelled by the end of 2027. It also reckoned only about 130 of thousands of "agentic" vendors are the real thing. It calls the rest "agent washing."

Raj Agent washing. Like my nephew saying he's studying because the laptop is open.

Aiman Then there's the bank-balance story from the start. Shane Mac runs a company called XMTP Labs. By his own account, he set up a Grok Bot agent called "CFO", as in chief financial officer. It was meant to send him a private monthly report. Instead, it posted his balances into the company's executive channel, after mixing up two chats with similar names. That rests on his own posts and essay, and hasn't been independently verified.

Mei Ling And on Meta's Muse, there's an admission from the top. In September, Vishal Shah, Meta's vice-president of AI products, said, "It is impossible to say that there is never going to be a mistake."

Raj At least he's honest. So if the bot sends the wrong invoice to the wrong customer…

Aiman Then you apologise, fix the rule, and ask whether that job should be the bot's at all. Now some history, because we've been here before. ATMs spread across America from the 1970s, and everyone assumed bank tellers were finished. Economist James Bessen found the average urban branch needed twenty tellers in 1988, and thirteen by 2004. But banks opened forty-three per cent more urban branches. Teller jobs didn't vanish. The job shifted towards solving customers' problems.

Raj Hmm. So the machine took the cash counting, and the human got the complaints.

Aiman Roughly. Factory robots are the harder story. Daron Acemoglu and Pascual Restrepo studied American robots from 1990 to 2007. Each extra robot per thousand workers cut local employment rates by roughly 0.2 to 0.3 percentage points. Wages fell too.

Raj So whether it's good news depends on who you are, and where you live.

Aiman That's the lesson. Then there's self-checkout. In 2024, CBS reported big American retailers scaling it back, over theft, scanning mistakes and unhappy customers. Dollar General removed it from 300 stores. Target limited it to ten items or fewer. One retail analyst, Neil Saunders, called it "a very love-hate technology."

Raj That's me and the supermarket self-checkout. It always says "unexpected item in the bagging area." The unexpected item is me. But here's my real point. For a Malaysian small and medium enterprise, the agent is the self-checkout. It looks cheap, then you hire someone to stand next to it.

Aiman I'd push back on that. Self-checkout moved the work onto customers. A good agent moves the boring work off your staff. That's different.

Raj Not if my staff spend Friday fixing what the bot did on Thursday.

Aiman Fair, I hadn't counted the clean-up time. That's a real cost, and most vendors leave it out of the brochure.

Raj I still print my invoices. So am I a dinosaur, or am I ahead?

Aiman You're the control group. Here's how I'd land it for small firms. On Monday, we costed a WhatsApp AI inbox. I said pay for the respond.io plan, at about RM649 a month. Raj said the free tools would do. Mei Ling wasn't sold. My vote stands. That's an agent with one narrow job and a human on call, which is the shape history rewards.

Raj Mine stands too. Start free lah, prove it works, then pay.

Mei Ling And mine stands. So we're split on the tool, but agreed on the shape. Pick one job where a mistake is cheap and easy to undo, and keep a person signing off.

Aiman A word now from the people who keep our lights on. After that, Mei Ling reports back from five other podcasts, and the part about Malaysia they all skipped.

Sponsor If your team keeps asking what AI actually means for their job, someone should give them a straight answer. AI-YA speaks at companies and events across Malaysia. Not a tools demo. A clear session on what's changing in your industry, what's hype, and what your people should do differently on Monday. Boardrooms, town halls, conferences, or a half-day workshop for the team who'll actually use it. Book at ai-ya.my.

Mei Ling Welcome back. I spent the week listening to other people's podcasts, so you wouldn't have to.

Raj Your weekend sounds very relaxing.

Mei Ling I took notes. Some of them have footnotes. Start with Hard Fork, from The New York Times, on the second of October. Reporters Mike Isaac, Erin Griffith and Eli Tan asked whether agents are "cute, cuddly and maybe catastrophically dangerous." The show set it against months of rogue agents causing havoc online. It also covered OpenAI and Meta launching personal assistants.

Aiman That's a callback for us. In our first episode, on the fifth, we covered OpenAI warning over a hundred organisations about its own rogue agents. Those were test agents escaping a lab. This week's worry is agents we invite in on purpose.

Mei Ling Second, the Big Technology Podcast. On the second of October, Alex Kantrowitz and Ranjan Roy of Margins covered the Dots launch, including Alex's own review. They asked whether a good assistant even needs a frontier model. A week later, on the ninth, the episode description lists Muse booking the wrong doctor. It also lists the Grok Bot leak of Shane Mac's banking details.

Raj The wrong doctor. I'd rather the bot leak my bank balance. It's a very short message.

Mei Ling Third, The Ezra Klein Show, also on the ninth. His guest was Princeton computer scientist Arvind Narayanan, who co-wrote "A.I. as Normal Technology" with Sayash Kapoor. Their case is that AI resembles past big technologies closely enough that we already have a road map. Fourth, The Diary of a CEO, again on the ninth. Steven Bartlett's guests, Nick Hanauer and Daniel Priestley, debated whether AI at work means mass unemployment or a new era of human potential. The show notes don't say who argued which side, so I won't either. Fifth, Bloomberg's Odd Lots, on the twenty-fifth of September. Joe Weisenthal and Tracy Alloway spoke to the writer Hayes Davenport about Hollywood. His argument was that things were starting to feel bad there before chatbots showed up.

Aiman Which matters. Not every job loss is an AI job loss. Our own minister made a similar point about business closures.

Mei Ling Now, what do they miss about Malaysia? I have six things. First, money. We're mostly small firms paying in US dollars. A hundred-dollar plan is a rounding error in San Francisco. Here it's about RM408, a fifth of the RM2,000 minimum wage.

Raj That's what I keep saying. Nobody on those shows asks what it costs in ringgit.

Mei Ling Second, access. Dots are rolling out in "eligible markets." Musk's bundle is priced for Americans. And our regulator has already restricted Grok once. San Francisco talks as if everyone gets the launch on day one.

Aiman I'm not convinced that's a disadvantage. Arriving later means the bank-balance mistakes happen to someone else first.

Mei Ling I hear what you're saying, but late adopters can get the job losses without the productivity gains. The ISIS data says exposed pockets here are already feeling it.

Aiman Let's split it, then. Wait for proof on the risky tools. Don't wait on the skills.

Mei Ling Agreed. Third, we're an export economy. TalentCorp's ten sectors make up sixty per cent of our gross domestic product, about RM933 billion. They include electrical and electronics, and global business services. That second one means back offices run here for foreign firms, the kind of desk work agents are being built for.

Raj Wait. So the bot in San Francisco could be doing a job in Cyberjaya.

Mei Ling It could. Which is why reskilling those teams matters more here than there. Fourth, language. Our customers write in Malay, English, Chinese and Tamil, sometimes in one sentence. A San Francisco demo doesn't test that.

Raj My customers mix three languages in one voice note. Good luck to the bot.

Mei Ling Fifth, data. A bot with access to your inbox holds your customers' personal data. Under the Personal Data Protection Act, that responsibility is yours. And Grok Bot's terms, as reported, put responsibility for an agent's actions on the customer. Sixth, the fresh grads. The podcasts debate whether companies still need juniors. Here, the ISIS data says early-career workers already carry more risk. So that's what a middle-income export economy has to ask. Not whether the bots arrive. Who pays for them, who checks them, and who still gets to learn. Now, the Friday positions. First, this week's Big Questions. Monday: should Malaysia own its AI infrastructure, or fund getting people to use AI?

Aiman Wednesday: does an employer hire the fresh graduate and push her to senior work sooner, or protect the entry-level jobs where people learn?

Raj Thursday: should public AI money back deep-tech champions at the top, or reach small firms at the bottom?

Mei Ling This week's question is one you can answer for yourself. When an AI agent can do a whole piece of your work, do you hand it over? Or do you keep a person signing off each step? Aiman.

Aiman Hand it over where mistakes are cheap and reversible. Drafting, sorting, reminders, first replies to customers. Start now, and measure the results. A person signs off anything touching money, personal data, or someone's livelihood. Waiting costs us the learning.

Raj I'm closer to signing off every step, for now. The tests show agents still finish only a minority of real projects. A small firm can't absorb a public mistake. Earn the trust task by task, then widen the bot's job.

Mei Ling I sit between you. The test is reversibility, not the technology. If you can undo it in five minutes, delegate it. If you can't, keep a human in the loop. And spend as much on people as on tools. Now the direction the country has set. The National AI Action Plan 2026 to 2030 rests on five pillars, from talent to digital trust. AI Malaysia Berhad replaced the National AI Office, and launched on the twenty-eighth of July. It houses a new AI safety institute.

Aiman AI Untuk Rakyat aims to reach 100,000 Malaysians aged eighteen to thirty by 2027. One of its six modules is called "Agentic AI for All." Budget 2027 adds another 100,000 free AI subscriptions for young people.

Raj And there's RM30 million for 4,000 small firms through MDEC, the Malaysia Digital Economy Corporation. Plus training through HRD Corp, the Human Resource Development Corporation, and the MSME digital grants.

Mei Ling But direction isn't delivery. Each of these starts as an announcement. Then it needs an allocation in a budget. Then a disbursement that reaches an actual person. We'll keep tracking all three.

Aiman So, what can you do? If you're a worker, move towards tasks where AI helps you, and get good at checking its work.

Raj If you run a small business, give an agent one job, keep a person signing off, and count the clean-up time honestly.

Mei Ling If you're a student, learn to direct and check agents, and still learn the basics underneath. Malaysia moves forward when each of us takes that next step. That's where the AI-YA team stands this week.

Aiman One thing to do this weekend. Take thirty minutes and write down one task you repeat every week. List every step. Mark a step green if a mistake there is easy to undo, and red if it isn't. Next week, try an AI tool on the green steps only, using dummy data. The red steps stay with you. And never put customers' personal data into a tool your company hasn't approved.

Raj AI-YA runs AI lectures, workshops and implementation training for Malaysian companies. Details at ai-ya.my/podcast.

Aiman Tomorrow is our Saturday week in review. We'll take you through the week's AI news for Malaysia. What moved since we covered it, and the biggest story we haven't covered yet. And the week's Verdicts and Big Questions, revisited.

Mei Ling And we want to hear your position. Would you hand a whole job to an agent, or keep signing off each step? Tell us in the comments on the episode page at podcast.ai-ya.my.

Raj If this was useful, follow the show and leave us a rating. It helps other Malaysians find us.

Mei Ling Rest properly this weekend, and give the people around you your full attention. Take care, and we will see you tomorrow for the week in review.