An Australian Energy Market Operator prediction of a seven-fold increase in power consumption in 10 years from AI use is the latest grim statistic to send shivers down the spines of legislators faced with a tidal wave of data centre proposals.

The forecast comes as Prime Minister Anthony Albanese takes a set of standards for AI deployment to the National Cabinet with data centres at the top of the list.

Energy Minister Chris Bowen says data centre operators have to supply their own energy needs from renewable sources. This edict is expected to be vehemently opposed by coal-friendly states such as Queensland. 

Small modular reactors and microreactors advocated by some big US tech firms such as Oracle co-founder Larry Ellison have been on the minds of some, but are not regarded as commercially ready technology yet.

A clash with states over coal is the least of Albanese’s worries. The biggest challenge will be garnering enough power for the AI datacentre juggernaut without it crushing domestic energy supplies.

And the cynics are circling. Successive Australian governments already have trouble getting big tech to pay its share of tax; they thumb their noses at government. Will they also thumb their noses at attempts to constrain data centre rollouts through “standards”? Can some projects be delayed to allow available energy supply to catch up? Can this massive industry be brought to heel?

Speaking at a Tech Leaders’ Conference at Bowral, Toby Walsh, Professor of Artificial Intelligence at University of New South Wales (UNSW) says it would be almost impossible to undo (or delay) many of the massively expensive projects already agreed to and now on the drawing board. So much money is at stake.

Professor Walsh also says that on current trajectories, renewables can’t be rolled out at the rates needed by data centres.The energy supply simply won’t keep up.

According to reports, data centre proposals have skyrocketed from 97 to 225 over the past year, a factor which saw Liberal MP Aaron Violi blast the Albanese government for being behind the eight-ball on implementing AI industry standards and regulations.

“Businesses, particularly the tech sector, expected a forward-looking roadmap and clear rules of engagement for the years ahead,” Mr Violi told the conference.

“Instead the government announced an office within an office or the Office of AI alongside a new set of regulations for data centres.

“To be clear we absolutely need a plan for data centres to ensure they work for our communities but a set of rules for data centres isn’t a substitute for a national AI plan and frankly it’s concerning that the Prime Minister and the government conflated the two.”

Mr Violi said parliament’s new Joint Standing Committee on AI, established a week ago, would seek to directly engage with the public on the AI rollout.

In his keynote speech, Assistant Minister for Science, Technology and the Digital Economy, Dr Andrew Charlton, said the government had three objectives: to capture the opportunity for  Australia with AI, to ensure all Australians share the benefits, and to protect Australians from the risks.

“The Prime Minister is taking those AI standards to National Cabinet this week with a view to legislating them in the not too distant future,” Dr Charlton said.

“The purpose of these standards is to make sure that we attract an outside share of global compute, but do it in a way that maintains social license, avoids the backlash that we’ve seen in other countries and delivers real tangible benefits to Australians.”

He says it is vital that Australians share the benefits of AI deployment “because I feel that if we don’t do this, we will lose social license very quickly … Australians more broadly have pretty low trust in AI.”

He says not only will data centres be required to “bring their own energy”, comprising gas, hydro and batteries, they will also have to pay their own transmission costs and contribute to grid stability.

“These requirements will not be voluntary. They will be mandatory.

“We’re not putting those rules onto these data centres in order to harm the industry, far from it. We are putting these rules onto data centres in order to put the industry on a stable foundation.”

Not everyone sees a future of massive AI-crunching data centres as the way forward.

Professor Walsh says it’s hard to see how investors in data centres will be rewarded with adequate returns.

“The estimate is that there’s going to be something like $7 trillion being invested into data centres around the world before 2030. It’s hard to understand how they’re going to see an adequate return on that investment.”

The fact data centres need to be refurbished periodically, and that GPU’s may last only 3 or so years before they are replaced, suggests that after ten or twenty years, they may end up as cold storage warehouses or distribution centres, he says.

Others at the conference spoke of a coming surge in cheap, on-device AI models that would see much big tech AI compute move onto devices.There is also signs of a pushback against large language model AI “hype” and “hysteria” and overinvestment.

IBRS industry analyst Dr Joseph Sweeney says Microsoft and big tech more generally thought they could gain a monopoly through large scale AI deployment. 

“They, meaning the big technology organizations out of the US, were thinking that they could do what Microsoft has always done: gain a new monopoly on the system.

“The problem was AI, or generative AI, or large language models, and more specifically vector databases, are an algorithm. You can’t copyright them, you can’t intellectually protect them in any way whatsoever, just like machine learning.

“And so when we saw this, and started writing this about three years ago, this slide, we were saying, yeah, this is going to be a disaster.

“What we are thinking is actually happening, that there has been a mass hysteria about how powerful and what AI actually is from the big tech bros.

“At least early last year, we could see that AI costs were going to skyrocket … they were losing money.

“Microsoft, about two years ago, saw this coming and pivoted. They didn’t stop building their own data centres, but they certainly postponed about $500 billion (worth) of them.

“Why would you do that? You do that because you realize that your overinvestment isn’t going to pay off, or maybe you’re not sure if it’s going to. So you’re hedging your bets. You’re pushing out that financial risk.”

SCX.AI CEO David Keane says that a year ago, everyone was using the proprietary big guys, but this is changing.

“Irrespective of where the information went, irrespective of our privacy and security and governance, we were using them.

“Now with local models that are running on these open weights environments, we can deliver cost savings and a significantly greater amount of control.”

Kiraa.ai founder Errol Brandt wants a moratorium on the building of data centres.

“I believe we need to stop building data centres in this country, we need to prove that the technology that’s already deployed is providing economic value before we start jacking up electricity prices in communities more than they already are, until we can sort out where our energy supply is going to come from. We haven’t worked that out yet.

“Number two, I don’t think that LLMs are the right technology. People are all putting their faith into token prediction machines.

“There are brand new architectures out there, which are much more suitable for reasoning and advanced decision making than simply predicting the next token.

“Number three, I think the whole cloud AI hyperscaler model, six of them, of the magnificent seven, the stupid six, have all gone down this path of capex.”

Published in iTWire, 25 August 2026 AI graphics courtesy of iTWire.

The post Experts fear big AI data centres will gobble up all our power. first appeared on Chris Griffith - Technology Journalist, Australia.

Buying AI tokens is costing a fortune. The penny has dropped that it’s much cheaper to use AI where it is really needed and stick to algorithms/traditional automation where it isn’t.

I do some development work on RAG chat bots and AI enabled business software as a sideline. When building workflows, you often have a choice of explicitly defining, say, a branch using traditional logic or have AI sort it all out for you. The former might take longer but over time it will save tens of thousand of dollars of token costs for clients.

Australian firm Workato says a token bill of $100,000 could be reduced to $10,000 if you only use AI in a workflow for where it’s needed. Another major saving is not to use AI models that far exceed the capability that you need. If you are developing a customer service chatbot, do you really need a LLM designed for deep research? The cost difference could be $0.30 per million tokens to, say, $10-50 per million for Claude Fable 5. A massive variation.

Look to Uber if you need a case study. Workato says the choice of model should be an agent by agent decision. There are other savings strategies. Nevertheless, keeping AI ongoing costs in check will be part of the new reality we face as production ramps up.

Don’t worry about bean counting; token counting may be the new preoccupation for many. Alex Zaharov-Reutt of iTWire and I recently sat down and explored this issue. Here is what he published in iTWire.

By Alex Zaharov-Reutt

Government’s AI pitch used to be all promise. On day 1 of Tech in Gov 2026 in Canberra, one of Australia’s most experienced technology reporters told me the promise has been replaced by something more useful: a hard conversation about what AI actually costs, who controls the data, and whether any of it can be secured.

I sat down for an iTWire TV interview at the close of day 1 of Tech in Gov 2026, the Terrapinn conference running 4 to 5 August, 2026 at the National Convention Centre in Canberra.

His read on 2026: the fantasy has drained out – and reality is in. 

“We’re seeing a bit of a movement away from the discussion that AI can do everything? Let’s get real,” Griffith told me. A couple of years ago the story was that agents would swallow whole processes of government and leave a trail of redundancies behind them. That has softened into something more measured, even as the job losses in tech keep coming, coders included.

What replaced the hype is money. And that, more than any keynote, is what the public servants were actually talking about.

The token counters are taking over from the bean counters

The session that clearly landed with Griffith was about token costs, the per-call price of running large language models. He has skin in this game. On the side he builds retrieval-augmented generation systems, bespoke applications that sit your own data on top of public models, so he has watched the meter run in real time.

His advice to anyone wiring up an agent is to use ordinary automation wherever ordinary automation will do, and to save the model for the parts that genuinely need it.

“You want to use automation, but not AI as much as you can, to save the cost of the calls to AI, because you can save a client thousands and thousands of dollars a month,” Griffith said. His worked example is stark: a badly built agent that reaches for the model at every step versus a well-built one that only calls it when it must. “It could be the difference between an agent costing $100,000 dollars a month and $10,000 dollars a month, if you do it correctly.”

The second lever is picking the right size model for the job. A customer-service chatbot or a basic financial calculation does not need the same frontier model as a research scientist working on quantum physics, and the newest models cost the most per token.

“You don’t need some agents to have the most state-of-the-art versions of AI, because the token cost is much, much more expensive. It’s horses for courses,” he said.

His line of the day, and the one I suspect gets quoted back at future conferences: “Instead of bean counters in government, we’re going to see token counters becoming the new discussion point in the public service.”

I have heard versions of this everywhere lately. There is the widely repeated (and never officially confirmed) story of a company that reportedly burned through half a billion dollars in token costs in a single month, and plenty of firms quietly admitting the models now cost them more than the people did. The whole promise was supposed to be savings. The bill says otherwise.

There is a familiar pattern underneath it. Griffith calls it the oldest trick in tech: “They lure you into a situation of dependency by making it free, very cheap. And then when you’re hooked, the price goes up.” He sees it as a developer too, where a platform that costs nothing at low volume quietly turns into hundreds of dollars a month once you lean on it.

Cost, more than capability, is now the tempering factor on AI. It is also why the efficiency race matters. Model makers including Anthropic have made token efficiency, getting more work done per token, a headline feature of their newest releases rather than a footnote, and you can see why in the pricing. Everyone gets that AI is powerful. If it bankrupts the department running it, that power is academic.

Data sovereignty and sovereign risk have been conference staples for a couple of years. Griffith’s sense is that government has actually improved here, and is taking it more seriously at a policy level.

The open question is enforcement. The government has moved on AI rules, including proposed mandatory guardrails for high-risk uses and a new Office of AI to coordinate the response. Griffith’s worry is what happens when that framework meets the biggest technology companies on earth.

“Are we going to get any better enforcement with AI than we have with tax?” he asked. His analogy is deliberate. Australia has struggled for years to close the tax workarounds that large multinationals engineer, and he expects the same cat-and-mouse dynamic to play out with AI.

“We’re going to struggle with our great new regulatory standards framework that the Prime Minister, Anthony Albanese, talked about,” he said. I noted that “Governments set up rules and regulations, and companies find their way around it.” Nobody at the conference seemed to disagree, which is its own kind of answer.

Cybersecurity in a dark room

If AI cost was the fresh anxiety, cybersecurity was the evergreen one. Griffith’s framing has a chill to it, because the threat has stopped being something defenders can fully picture.

Put AI in the hands of scammers who understand it, add the quantum computing that is coming, and you get intrusions that defenders may not even be able to imagine, let alone stop.

“You’re walking around in a dark room without being able to see the walls,” he said.

He watched representatives from the Department of the Prime Minister and Cabinet and Parliamentary Services describe getting ready for attacks at that level, including running simulations to predict what might come. His takeaway was that nobody has a clean solution yet.

That includes the vendors. Griffith singled out a presentation from HP on a threat that keeps security people up at night: not just the poisoning of data, but the poisoning of AI agents themselves. If an attacker can reach the instruction set you give an agent and quietly rewrite it, they can turn your own automation against your data. It is the kind of risk catalogued in the OWASP Top 10 for large language model applications, and it is no longer theoretical.

“We can actually turn AI agents around and make them absolute enemies of the companies that employ them,” Griffith said. His conclusion, which I concurred with, is that the defence is more AI, not less: “We’re going to need AI agents that are good and on our side and unhacked, to protect us from the ones that are malicious.

The moment from the conference floor that clearly rattled Griffith came from the CSIRO. He relayed a warning from a CSIRO speaker that cuts against a decade of identity design (I could not pin down which CSIRO presenter said it, and I would welcome the correction if anyone at the session can name them).

The line, as Griffith recalled it: “A biometric signal is no longer independent proof of identity.”

Sit with that for a second, because we have built a lot on the opposite assumption. Face, fingerprint, voice: we treat each as its own proof. The message from CSIRO, he said, is that the only reliable path now is a multi-layered approach, several signals combined, because any single one can be faked. It echoes what identity researchers globally have started saying out loud, that no single biometric wins trust on its own anymore.

Two-factor authentication was step 1. Griffith’s point is that real multi-factor, multi-signal identity is now the baseline.

Detection is getting harder in parallel. He noted the same CSIRO framing on video: it is becoming much harder to be sure a clip is not a deepfake. Layer that on top of an internet that many people already treat as their source of truth, and the confusion compounds.

He tied it to a media story close to home. With the news media bargaining arrangements being reworked (the government has floated a levy on Meta, Google and TikTok to fund journalism), Griffith worries about what happens if trusted news weakens while social platforms fill the gap with synthetic content.

“We see how much AI slop is out there, and with a little bit of tweaking, that can be much less slop and much more convincing fakery,” he said. “It’s a pivotal time for us to get this right.”

Shadow AI, and the agent nobody remembers switching on

On the defensive side, there was a note of cooperation. Griffith pointed to the National Cyber Security Coordinator, Lieutenant General Michelle McGuinness, who described a rising level of cooperation across the tiers of government and new partnerships aimed at a unified response to threats against infrastructure and defence.

Then there is the threat from inside, which Griffith summed up in two words: shadow AI. In the public service, plenty of AI use is sanctioned. The danger is the staffer who reaches for an unsanctioned tool without realising that the act itself can leak information out of a department, or open a door for someone to get in.

The part that should worry any CISO is the time lag. “The problem is that you may not know for a long time that it’s occurred,” he said. Worse, an agent can be stood up, do its job, and then simply keep running after everyone has forgotten it exists.

“After some time there is no realisation that the agent is still operating there, and who knows what could be happening with it,” Griffith said.

This is exactly the discovery problem the identity industry has been circling, and the vendors are now hunting for it. Griffith described tools that scan an organisation for every AI agent present and routinely turn up dozens that were never sanctioned, which then get shut down. It maps onto the pitch I have heard from identity players like Okta, and onto a wave of shadow-AI discovery features landing across the governance market.

His logic is hard to argue with: “They can’t shut it down if they don’t even know it’s there, and they don’t know what it’s up to.”

Ghost Bat, Ghost Shark and a stockpile of drones

Defence brought the conversation somewhere more physical. The Department of Defence talked up further AI in the field, and Griffith zeroed in on two Australian programs with excellent names.

The Boeing MQ-28 Ghost Bat is the pricey one, an autonomous “loyal wingman” drone that can range across the country on surveillance runs and is heading toward combat service later this decade. The Anduril Ghost Shark is the quieter one, an extra-large autonomous submarine that can stay under for weeks, with the first units already delivered to the Royal Australian Navy and a manufacturing line up in Sydney.

For a country Australia’s size, Griffith argues, the endurance is the whole point. We need long range and long time-on-station more than most nations do, in the air and under the water both.

The lesson he keeps returning to comes from Ukraine, where cheap autonomous systems have rewritten naval and air warfare, from the Magura sea drones that pushed Russia’s Black Sea Fleet back to the deep strikes on Russian bombers. “Ukraine has shown us how important drones are,” I noted.

His caveat is the security one, and it connects straight back to the cyber conversation. A drone fleet is only an asset if it stays yours. “We need them autonomous, and we need them secured. They can’t be taken over by a nation state and turned against us.”

Griffith’s last stop was the quiet engine room of government data. He spent time with the team behind the Digital Atlas of Australia, the Geoscience Australia platform that pulls trusted national datasets onto a single interactive map and checks that government information, mapping and Australian Bureau of Statistics data included, is fit to be used.

It is timely work, because the 2026 Census lands on 11 August, a week after we spoke. Griffith has his envelope already.

He asked the obvious question: are they using AI on all this? The answer, for now, is not yet.

“We’re looking at it, but we haven’t taken that big step yet to actually implement it,” he was told. I made the observation that the ABS has managed the census results for 100 years without it, and is in no rush to graft it on before it is ready.

Griffith thinks that will change, and soon. Once you add satellite imagery and mapping to the mix, the ability of AI to read visual data at scale becomes too useful to leave on the shelf.

“AI can slice and dice even more quickly than traditional methods,” I added as you’ll see in the video interview, noting “we really can harness this granular data in ways that were just not humanly possible before, without taking way too much time.” The holdup now is about comfort and trust; the capability is already there.

The bottom line

What struck me most, listening to Griffith at the end of day 1, was how unglamorous the smart version of this has become.

The attitude has shifted hard from 2 years ago, when we all started talking about agentic AI as though it would run everything by itself. The 2026 conversation is granular: work out where in a workflow AI genuinely earns its keep, use plain automation for the rest because it is cheaper and more reliable, and treat every token, every agent and every biometric as something you have to account for.

“Automation is used where it’s needed, and AI is used where it really can make a difference,” Griffith said. That is a more mature idea than “AI, do it,” and a much cheaper one.

He is more confident than he was, though not relaxed. Governance is still the hard part, enforcement still an open question, and the security clock is still running against defenders. And the models keep getting stronger.

“It’s going to be frighteningly more powerful as we go along,” he said, half-smiling. “Maybe the science fiction films will come true.”

This article is based on Alex Zaharov-Reutt’s iTWire TV interview with Chris Griffith at Tech in Gov 2026, held at the National Convention Centre, Canberra, on 4 August 2026. Griffith is a former senior technology writer at The Australian, an iTWire contributor, and the publisher of chrisgriffith.com. Some quotes have been lightly trimmed for length and clarity.

Originally published in iTWire, Aug 5, 2026. AI graphics courtesy of iTWire.

The post From bean counters to token counters: how government overuse of AI could cost billions. first appeared on Chris Griffith - Technology Journalist, Australia.

Personal finance and trading education platform RationalFX has compiled layoff data from verified sources that show the reach of retrenchments already in 2026.

“Mounting warnings from business leaders and economists point to artificial intelligence as a key accelerator of these layoff waves, with companies restructuring around automation, machine learning, and efficiency gains putting not only individual roles but entire job functions at risk,” RationalFX Data Analyst Alan Cohen warns.

He says data shows that more than half of the 30,700 tech layoffs worldwide since the start of the year have come from a single company: Amazon.

“The US tech giant announced 16,000 cuts in early 2026, following 14,000 roles shed in October 2025. That earlier round made Amazon the second-largest contributor to global tech layoffs in 2025, with a total of 19,555, just behind Nvidia’s massive 33,900 job cuts.”

RationalFX has ranked the top 10 companies responsible for layoffs in 2026 so far: Amazon – 16,000 layoffs, ams OSRAM – 2,000 layoffs, Ericsson – 1,900 layoffs, ASML – 1,700 layoffs, Meta – 1,500 layoffs, Block – 1,100 layoffs, Autodesk – 1,000 layoffs, Salesforce – 1,000 layoffs, Ocado – 1,000 layoffs, and Pinterest – 677 layoffs.

Country wise, the US leads tech job losses with 24,600 followed by Sweden 1900 and The Netherlands 1700. Tech powerhouse India has suffered 920 layoffs this year so far.

Mr Cohen says large-scale layoffs were once considered a red flag by investors, but they have become a standard tool for operational refinement among leading tech firms.

“Amazon’s massive layoff waves clearly illustrate this shift: even as the company posts record revenues and pours billions into AI infrastructure, it is flattening management layers and eliminating entire job functions,” Mr Cohen says.

“CEO Andy Jassy’s vision frames this not as contraction, but as operating ‘like the world’s biggest startup’, nimble, AI-augmented, and future-proof.

“Across Europe, ams Osram’s 2,000 layoffs follow a similar approach: the company is consolidating global operations despite improved financial results, reflecting a broader shift towards efficiency-led restructuring.

“Going forward, 2026 appears set to be a year in which layoffs are deployed not as a signal of financial distress but as a deliberate lever to sharpen competitiveness and focus investment on high-return priorities.

“To determine which companies led 2026’s biggest job cuts, the team at RationalFX compiled layoff data from multiple verified sources, including U.S. WARN notices, TrueUp, TechCrunch, and the Layoffs.fyi tracker, covering announcements made since the start of 2026. ”

RationalFX’s full report can be viewed here.

Published by Channel News Australia, February 19, 2026

The post AI sparks massive layoffs In the profitable tech sector first appeared on Chris Griffith - Technology Journalist, Australia.

Until recently Netflix believed it was in the box seat, saying “the only signed, board-recommended agreement with Warner Bros is the only certain path to delivering value to Warner Bros’ stockholders.

All seemed on track. Netflix was prepared to pay $27.75 a share in cash to buy Warner’s studio and streaming business in a deal worth $72bn. In addition, Netflix CEO Ted Sarandos and Warner Bros CEO David Zaslav had forged a good relationship, and there was confidence in Netflix’s ability to take on the challenges.

Paramount on the other hand sought to buy the entire Warner Bros business including its cable business which included CNN, Discovery and TNT. They carried substantial debt.

So Netflix had the running. In the minds of some investors, the Paramount deal was unrealistic with its valuation of Warner Bros at $30-31 per share. There was a major antitrust concern about the deal. Netflix too had the ability to match any competing Paramount offer.

But there were other issues at play. Importantly, the Trump administration saw Paramount Skydance as a better option politically. A Paramount Deal would put Warner Bros in more conservative hands with Oracle founder and CTO Larry Ellison, the father of Paramount CEO David Ellison, a long time Trump ally and financier.

Some saw the benefit of a friendlier Paramount having ultimate oversight of CNN which had been hostile to the Trump administration.

The New York Post reported that Paramount had begun legal action against Warner Bros claiming that it had been unfairly treated due to the friendship between Netflix CEO Ted Sarandas and WBD CEO David Zaslav. Paramount itself mounted a hostile offer to buy WBD at the end of December.

At the same time, the Netflix camp was concerned that Warner Bros would be left with a debt-ridden orphan in its cable channels whose sale wouldn’t help Netflix finance an offer above $30 per share to match Paramount’s.

Warner Bros this week told the Post that it still favoured the Netflix deal and planned a shareholder meeting on March 20.

“We continue to believe the Netflix merger is in the best interests of Warner Bros shareholders due to the tremendous value it provides, our clear path to achieve regulatory approval and the transaction’s protections for shareholders against downside risk,” says WBD Chairman Samuel A. Di Piazza Jr, the Post reported.

But the tide shifted. There was new opposition within Warner Bros to the deal with The Wall Street Journal reporting that investor Ancora Holdings planned to oppose it, saying it is underdone.

Ancora, which has an almost $200 million stake in Warner Bros, told The Wall Street Journal the Netflix board had not engaged enough with WBD.

Simultaneously, Warner Bros CEO David Zaslav had left the door open to a new bid from Paramount.

The New York Post reports that while Zaslav had good relations with Netflix’s Sarandas and supported its deal publicly, he also was keen to keep the bidding war going. Zaslav wanted to fan the flames of competition for a better share price.

“I wanted to put these guys in the ring together and let them duke it out,” Zaslav told one person close to the matter, the Post reported.

The outcome is that Warner Bros will restart talks with Paramount in what will be a rekindled bidding war. For its part Netflix says it will grant WBD a seven-day waiver to negotiate with Paramount, which has until February 23 to submit a best offer.

Netflix then has a right to match the offer, but whether Netflix shareholders agree to a heftier share price remains to be seen. Netflix says it still has a deal with Warner Bros.

Published in Channel News Australia, 18 February 2026

The post New war breaks out In the battle to own Warner Bros first appeared on Chris Griffith - Technology Journalist, Australia.

The company has introduced integrated artificial intelligence, giving everyday users ways to troubleshoot and fix their phones, tablets, and a wide range of gadgets at home or on the go.

In a media statement today, iFixit announced its new AI tool, FixBot, which lets a person describe their problem by voice or text. “It does what a good expert does: asks the right questions, narrows down the possibilities, and guides you toward the fix,” says iFixit CEO Kyle Wiens.

“You tell it what’s happening: your phone dies at 30 per cent, your washing machine won’t drain, your mower sputters and stalls.

Wiens says FixBot actually knows what it’s talking about.

“It pulls its answers from our 125k repair guides, our massive question-answer forum, and our huge cache of PDF manuals.

“It knows how to find ideal bolt torque from a table in a manual. It can read a part schematic and tell you the part number you need to order. Importantly, it’s much better than other systems at not making stuff up. It’s not going to tell you to cut the blue wire when there is no blue wire.”

Alongside FixBot, iFixit has also launched the iFixit app, available through the App Store and Google Play Store. Wiens describes it as “a workbench that keeps track of your repairs.” The app also includes a battery lifespan predictor.

iFixit’s history with apps is unusual. It first launched one in 2011, but in 2015 it was removed from the App Store after it published a “tear‑down” of an Apple TV developer unit. Despite that setback, the company continued to thrive through its website. Now, ten years later, the app has returned, featuring thousands of repair guides, smart troubleshooting, and support from FixBot.

Founded in 2003, iFixit today offers nearly 69,000 repair guides covering everything from smartphones and laptops to cars, appliances, and even medical equipment. The company is also a passionate supporter of the right‑to‑repair movement and generates income by selling parts and repair tools.

Published by ChannelNews Australia, 11 Dec 2025

The post iFixit puts an AI mechanic in your pocket first appeared on Chris Griffith - Technology Journalist, Australia.

Measures to block under-16s from accessing popular social media platforms are proving to be controversial and, so far, imperfect.

On the face of it, Australia’s under-16 social media ban has proved popular with voters. Some 77 per cent of Australians supported the ban in a YouGov poll in November 2024, while another poll by media and marketing specialist B&T taken two months later registered 68 per cent parental support. Teenagers understandably don’t favour it, but in a survey by consumer insights consultancy Nature, two-thirds acknowledged the toxic nature of certain social media platforms.

At its worst, the internet can present as a veritable cesspit of cyberbullying, rape threats, grooming by pedophiles, sex-related extortion, deepfake image-based abuse, misinformation and, more recently, unhealthy attachments to AI chatbots. For its part, the Federal Government has chosen to focus on selected sites that are particularly attractive to millions of young Australians: Facebook, Instagram, X, YouTube, TikTok and Snapchat.

In practice, the implementation of the ban has been a convoluted process. More than 50 companies took part in a trial that involved either verifying users’ age or providing an assurance of a user’s age. The verification options included government ID checks, such as passports and licences, mobile carrier data, and biometric methods such as card-based checking. Standalone biometrics that could be used without storing data, and methods including face recognition, voice analysis and hand movement analysis, all of which are used to estimate a user’s age.

The Age Assurance Technology Trial’s preliminary report was released in June and its general conclusion was encouraging enough. “Age assurance can be done in Australia and can be private, robust and effective,” it said. But a significant issue emerged with the subsequent resignation of a member of the trial’s overseeing panel. Tim Levy, managing director of Qoria, one of about a dozen Australian companies taking part, questioned its independence.

Some of the technologies tested had a reported 85 per cent error rate in estimating the age of users, requiring “18-month tolerances” and possibly multiple age-assurance steps. Levy says: “These results were predicted and echo sentiments expressed by Qoria and the parental control, school safety and social media platforms last year.”

According to him, UK-based Age Check Certification Scheme (ACCS), which oversaw the trial, has created an age assurance industry. “They’ve set the standards for age assurance and then they’ve gone and tested it,” he says. “They wanted to prove that it works, because that’s their business.”

Levy is not confident age verification alone can be effective. He believes the government’s focus on a handful of popular platforms will protect children from only one corner of the internet, and that mobile device management systems that can filter content across the internet would achieve a better outcome.

“The power of these tools that businesses and big schools, mostly in the US, have, and private schools in Australia have, is astounding,” Levy continues. “They can provide a safe, age-appropriate experience across the entirety of the internet, just not on the six social media platforms.”

Some, such as the Age Verification Providers Association, a global trade body for providers of age assurance technologies, believe there are simpler methods than biometrics for accurate assessments. “If anyone tries to claim they have devised an algorithm that can accurately assess your age within a day, a week or even a month of your real age based solely on a few selfies, they are not being truthful,” the association said in a media release in June.

Simpler methods could include utilising GovID or one of the ID verification technologies being developed by banks, telcos and other service providers. Another suggestion is for users to obtain an encrypted QR code from their state’s office of Births, Deaths and Marriages to be provided as proof of age to social media platforms.

The Federal Government’s COVIDSafe app was a prime case of sticking too long to an advanced but flawed technology solution, as low-tech QR verification ultimately proved more practical for tracking infections during the pandemic.

Ric Richardson, the Australian inventor who patented anti-piracy software for product activations, recommends taking government checks out of the equation altogether. Instead, he says, parents can add their children’s phone numbers to a blacklist. His SafeGen system would require social media companies to check this register when performing verifications. “Parents know exactly how old their kids are, and they know exactly what devices their kids use,” he says.

About a dozen Australian firms, including Sydney-based Australian Payments Plus, which showcased its established identity verification tool ConnectID, were among the 53 in the Age Assurance Technology Trial. “When a business requests age verification from their customer, ConnectID allows that user to choose a trusted entity,” says Andrew Black, the company’s managing director. “For example, their bank can confirm whether the user is over the required age, returning a simple ‘yes’ or ‘no’ response without needing to share date of birth with either the business or ConnectID.”

Other Australian companies in the trial included FrankieOne, Deep Media, One Click Group, R2 Labs, RightCrowd, ShareRing and TomorrowX. FrankieOne co-founder and chief technology officer Aaron Chipper says his company has aggregated more than 350 verification tools. “We can bring those different technologies together for the social media companies to be able to pick and choose what’s most appropriate for the customer,” he says.

Technology aside, there is also a risk that labelling the initiative as a ban will trigger FOMO (fear of missing out) in young teens who may seek to circumvent age verification via alternative strategies, including using a VPN or registering accounts in other countries. Australia’s eSafety Commissioner Julie Inman Grant is aware of this. At her National Press Club address in June, she reframed the ban as more of a delaying tactic, offering a timeout from the internet’s most toxic platforms during which teens can be taught how to identify and avoid dangers, as well as identify misinformation.

“Calling it a ban misunderstands its core purpose and the opportunity it presents,” she said. “We are not building a great Australian firewall … it may be more accurate to frame this as a social media delay, giving children a reprieve from the persuasive pull of platforms engineered to keep them digitally entranced – and entrenched.”

The education options for under-16 teens, however, have not received the same analysis and debate as the technology proposed for age assurance. These include a curriculum packaged by the Australian Curriculum, Assessment and Reporting Authority (ACARA) about the dangers of online platforms suitable for young teens, while the eSafety Commission has also compiled a website dedicated to online safety.

Others want to see measures beyond banning under-16s from social media platforms. Australian Martin Dougiamas is the founder of Moodle, a free open-source learning management platform he says is used by two-thirds of the world’s universities. Organisations download and run the platform on their servers to create a learning management system. He says educating children to navigate toxic forms of social media was like “teaching them to run safely down the middle of a highway”.

“We’re not actually attacking the real problem, which is that the platforms themselves are causing the problems,” he says. “The fact is they’re owned and controlled by profit-focused companies who are using advertising models. I would rather see the government put energy into Australia building its own infrastructure. Anybody in Australia can make new platforms on the internet. It’s something we just need to decide we can do, and that we’re not a second-class citizen in the world that relies on American companies to provide our infrastructure.”

Western Australian-developed DiGii Social, New Zealand’s MyMahi and German-based federated network Mastodon offer social media alternatives where standards can be set and enforced locally.

The ban, however, is far from the only shot in Inman Grant’s locker. She has worked with tech companies on an online code to limit children’s access to pornography, violent content, suicide and self-harm themes and sites that encourage disordered eating. Her office is driving a measure that will prompt an age assurance check for anyone logging into a search engine account, and the commission has also formed a youth council to liaise with teachers, parents, carers and young people. Yet still, it’s the proposed social media ban that continues to capture the most headlines.

Originally published in The List, Innovators 2025, The Australian, October 10, 2025

The post Screening teens for the under 16 social media ban first appeared on Chris Griffith - Technology Journalist, Australia.

While this is an initial dialogue, the MEAA has described it as “a breakthrough.”

“A year after launching our campaign to force Big Tech giants to stop stealing the work of Australia’s media and creative workers, the tech industry has agreed to come to the table and negotiate a compensation deal,” MEAA said in a statement.

“We look forward to sitting down with the Tech Council of Australia and the ACTU to reach a deal that serves our members’ interests swiftly and fairly. MEAA welcomes the acknowledgment by technology companies that Australia’s media and creative workers deserve to be paid for the content used to develop highly profitable AI models.”

The Tech Council appears receptive. CEO Damian Kassabgi noted that tech platforms have helped creatives gain exposure, clicks, and new revenue streams—including international reach for Australian artists.

“We’re keen to explore what AI means for different sectors, including opt-out models,” he said. “There’s been real progress in recent days toward a shared understanding of the need to collaborate and seize the national opportunity AI presents. This is just one part of that.

“We’re hopeful we can find a path forward on copyright that enables AI training in Australia while protecting the livelihoods of creators. What that path looks like has not yet been determined.”

This willingness to talk marks a major shift. Until now, big tech has warned that Australia risks falling behind unless our writers, artists, composers, photographers, researchers, and news organisations surrender their copyright to AI systems that consume their work—without consent or compensation. Many suspect the real goal is simply to use creative works for free.

This development also represents a reversal by the Tech Council, whose newly appointed chair, Atlassian co-founder Scott Farquhar, had previously urged Australian creatives to forgo copyright. His comments were disappointing, especially given the Council’s central role in shaping Australia’s tech economy.

Decades ago, I worked in the TAFE system and saw firsthand the value of collaborating with government to forecast employment needs and align education and industry. That’s precisely the kind of strategic foresight the Tech Council seeks to offer today.

The Council’s membership includes major Australian and U.S. tech firms, making its position especially influential. These companies are among the largest providers of high-tech jobs in Australia and must be part of any forward planning.

The Productivity Commission also weighed in, though its recent media remarks on issues including copyright felt more like thought bubbles than considered analysis. The claim that paying creatives would collapse productivity and ruin the economy was patently absurd.

Speak to those involved in AI governance and you’ll hear a different story. They’ll tell you that AI development won’t be impeded by copyright claims. They’ll explain that copyrighted material can be logged as it’s ingested into AI systems. The sky won’t fall if creatives assert their rights.

In fact, compensation claims are feasible, especially at the point of original use by AI. A log could form the basis of a searchable register, maintained by the Copyright Agency or another authority, where creatives could lodge royalty claims.

Yes, complexities exist, particularly with visual media. Legal disputes may arise over whether an image contains unlawfully copied elements. For news articles, direct negotiation between publishers and AI firms (akin to the News Media Bargaining Code) would be more efficient than tracking individual stories.

Importantly, many cutting-edge industries will rely on bespoke sovereign LLMs trained on proprietary business and industry data and not necessarily on creative works. At the recent TechLeaders 2025 conference, Australian company Maincode briefed journalists on its sovereign LLM strategy, which may not require local creative content at all.

Governance experts agree: technology itself doesn’t prevent us from paying creatives. What does? Greed.

Further, these discussions could lead to compensation models outside the Copyright Act. Historically, AI-related copyright disputes have played out under U.S. law, where weak “fair use” provisions have left creatives shortchanged. But Australian law is poised to take a more active role, especially as locally developed LLMs train on domestic material. Representative organisations are urging the government to address AI-related copyright explicitly.

There’s clear logic in updating the Copyright Act to reflect the realities of AI training. The Act was never designed with AI in mind, and without reform, we risk wasting millions on litigation to clarify ambiguous law. The US has already endured costly, protracted battles over whether LLM training breaches copyright. Big Tech denies it outright—and so far, U.S. courts have sided with them.

Adding provisions to Australian law that affirm copyright protections in the context of AI training would eliminate this ambiguity. Yet Industry Minister Tim Ayres recently reiterated that the government has no plans to amend the Act in either direction.

Meanwhile, some U.S. companies may still face sanctions for using pirated content—Anthropic, developer of Claude AI, being one example.

The irony? AI itself shows it’s not hard to devise ways to pay creatives. I recently brainstormed this issue with Microsoft Copilot. Together, we explored a model that logs works as they’re used to train LLMs, feeding them into a searchable register for royalty claims across various use categories. The dialogue was illuminating – see below. 

Of course, Australian organisations representing creatives have already proposed legal solutions. I defer to their expertise entirely.

Originally published by iTWire, August 22, 2025

The post Big tech to discuss compensation with creatives for AI training  first appeared on Chris Griffith - Technology Journalist, Australia.

Speaking as chair of the Tech Council of Australia—which represents both local startups and global tech giants operating in the Australian market—he outlined opportunities ranging from digital embassies and expanded API access to government services, to the potential for low-cost data centres powered by abundant green energy.

Much of Scott Farquhar’s vision was inspiring. But some elements deserve scrutiny.

One of the more contentious claims was his suggestion that Australian creatives and businesses should relinquish copyright protections when their intellectual property is used to train AI large language models. Farquhar argued that insisting on copyright claims could leave Australia lagging behind other nations, and advocated for relaxing local copyright law in line with jurisdictions like the United States.

In essence, he presented a binary choice: either forgo your copyright claim, or risk undermining the future of Australia’s AI industry. The alternative—preventing AI companies from crawling your site for content—was framed as equally problematic.

But there is a third path: requiring AI developers to pay a reasonable fee for using intellectual property. Big tech will still access the Australian content it needs. The sky won’t fall in, and the local AI sector won’t collapse.

Farquhar isn’t alone in this view. The Productivity Commission and others have echoed similar positions. While the Albanese government has so far resisted efforts to dilute copyright protections, its initial response in the previous parliament was underwhelming—convening a reference group and commissioning limited research, with little tangible progress.

Yet reform is urgently needed. Current copyright law was not designed with AI in mind, and updating it is both reasonable and necessary. The most constructive path forward would be to introduce specific provisions in the Copyright Act that address how AI models use intellectual property.

Big tech firms have long argued that training AI models—particularly large language models—is fundamentally different from piracy or unauthorized reproduction, and doesn’t affect creators’ income. Former UK Deputy Prime Minister Nick Clegg, speaking after his tenure at Meta, claimed that even requiring permission from creators would “instantly kill” the AI industry in a country.

US courts have largely accepted the distinction between training and reproduction. However, a recent judgment against AI developer Anthropic homed in on the use of pirated material—highlighting the issue of acquiring content illegally, rather than the boundaries of copyright law itself.

Australian courts may take a different view, given our stronger codified fair dealing provisions. But rather than leaving the issue to litigation, the government should proactively amend the Copyright Act to clarify AI’s use of IP.

A tiered payment schedule could be introduced: one for works used solely to train models and then discarded, another for content that’s retained and reproduced in some form. This would remove ambiguity and reduce the risk of drawn-out legal battles.

AI developers would still gain access to the content they need—while fairly compensating creators. 

Farquhar also proposed the creation of “digital embassies”—data centres on Australian soil that operate under the laws of client countries. This could appeal to nations across Asia, offering a secure and sovereign data solution. But it raises important questions: Could foreign governments sidestep Australian copyright law when training LLMs here? Would Australian law enforcement be hampered if such centres were used for espionage?

Despite these concerns, the concept is worth exploring.

Farquhar also emphasized the importance of developing AI to safeguard national sovereignty. He said this doesn’t necessarily mean building LLMs from scratch. Not everyone accepts this. You’d need to go back far enough in the AI supply chain to safeguard independence.

Bias in AI is now a pressing issue. In the US, President Trump has called for the removal of climate change, diversity, equity, and inclusion content from federal AI models. Another White House executive order promotes the global adoption of US-developed AI.

Whether Australian LLMs built on US foundations will inherit those biases remains to be seen. But it’s reasonable to expect that Australian models should reflect Australian values—especially where bias is concerned.

There’s an even darker dimension. Authoritarian regimes have long understood the power of controlling populations through media and social platforms. AI amplifies that control.

It’s a path humanity must not take.

Originally published by iTWire, 11 August 2025

The post Scott Farquhar’s AI copyright dilemma first appeared on Chris Griffith - Technology Journalist, Australia.

Its offerings will create targeted marketing strategies based on your company data, and use those strategies to devise and build personalised ad campaigns, including videos with your company branding.

The software giant is promoting 10 new AI-enhanced “agents” to its marketing platform customers at its Australian summit this week.

Workday, Oracle, Salesforce, Amazon, Google, Microsoft, OpenAI, Nvidea and SAP are others offering AI agent solutions.

AI agents are generative AI equipped components that can follow instructions and complete a set of tasks without needing continual human guidance.

In the world of the near future, AI agents used by event organising companies, hotels and transport services might collaborate to organise a conference, including accommodation and transport schedules.

Adobe’s agents cover a large range of marketing tasks, such as data insights and analysis, audience analysis, content production, data management, building sales pipelines, hypothesising and simulating ideas, customer journey optimisation, and workflow optimisation.

The agents were announced in March but the rollout is yet to begin in earnest in Australia.

Head of Product Marketing (APAC) Adobe Jeremy Wood said Adobe was collaborating locally with Marriott to demonstrate how its agents worked.

Mr Wood said with traditional data insights, marketers tried to figure out the performance of their respective marketing activity.

“A lot of the time pulling that data is quite a complex process. We have marketers saying I have to queue that up with the data analyst team, the data analyst team might come back to me within days, might come back with me in weeks.

“Now with agents, that agent can automatically go pull that data, bring it back and visualize it for anybody.”

An agent might also make recommendations about optimisation or nominate campaign strategies that work better. “This really makes it (analytics) accessible to many, many more users.”

Mr Wood said you might ask Adobe’s production agent to create content and imagery for an audience of a certain age group that likes ocean activities and being by the sea.

The agent would use both your branded content and film clips from Adobe Stock to build targeted marketing videos in more than 30 languages.

“We have a great case study with Coca-Cola, who’s done all of this exactly.”

He said customers were free to train their agents on their data.

An AI powered design system co-developed by The Coca-Cola Company and Adobe lets designers train AI models to achieve accurate content creation and quality.

Adobe’s agents are built using a platform called Adobe Agent Orchestrator and can work with third-party ecosystems. Adobe and Microsoft have jointly worked to enable Adobe Marketing to collaborate with Microsoft Copilot. Customers can build and train their agents, and agents might combine skills to complete certain tasks.

Adobe had also developed a consumer facing tool called Brand Concierge, to enhance a company’s communication with its own customers.

There is the elephant in the room; the fear of massive job losses from AI, and specifically agents. For data analytics, that could be the jobs of data scientists.

Like many in this field, Mr Wood is optimistic that AI agents will free up time for higher level strategic thinking.

“We’re seeing … on the enterprise marketing side … that this will free up some very valuable and skilled time. Those employees (will) be able to focus on much higher impact and much more strategically valuable exercises.”

Originally published by The Australian, 7 July 2025

The post Adobe AI to transform the way consumers see ads first appeared on Chris Griffith - Technology Journalist, Australia.

Technology publication The Verge has raised this prospect after examining changes to the website promoting the T1 8002 Gold Version.

“One of the phone’s main selling points was that it was to be made in America,” says The Verge.

“We figured that was unlikely to be true. And we were right: sometime in the last several days, the Trump Mobile site appears to have been scrubbed of all language indicating the phone is to be made in the USA.”

It says a huge banner on the homepage that says the T1 is MADE IN THE USA has gone, to name one example.

Indeed our examination of the website promoting the T1 says it is “proudly American”, “Designed with American values in mind”, has an “American-Proud Design”, has “American hands behind every device”. But today, when we looked, there is no explicit mention of the phone being made in America.

A footnote to the promotion site says products and services “are not designed, developed, manufactured, distributed or sold by The Trump Organization or any of their respective affiliates or principals”.

“T1 Mobile LLC uses the TRUMP name and trademark pursuant to the terms of a limited license agreement which may be terminated or revoked according to its terms.”

T1 Mobile appears to be a company associated with President Trump’s sons Donald Jr and Eric.

Channel News Australia has reached out to The White House for clarification.

According to its Wikipedia entry, T1 Mobile was launched in June 2025 (this month) as a venture by Donald Trump Jr and Eric Trump and is a licensed brand of the Trump organisation. It was launched on June 16 which was the 10th anniversary of the announcement of Trump’s 2016 presidential campaign.

The Wikipedia entry said Trump’s two sons announced the smartphones would be “exclusively manufactured within the United States with the goal to encourage other multi-industry companies to manufacture their products in the US”.

However, the Wikipedia entry pointed to analysts who said there wasn’t fabrication plants and factories that would enable the phone to be made within the country.

Earlier versions of the Trump T1 phone included a signature MADE IN THE USA claim which is now missing. Picture: Getty Images

The Huffington Post quoted Tinglong Dai, a global supply chains expert at Johns Hopkins Carey Business School, who said: “This phone is essentially a licensing deal.”

“The Trump Organization doesn’t design, manufacture or sell it; they’ve simply lent their name,” he explained. “That makes the idea of leading a US-based manufacturing effort even less credible.”

CNN meanwhile says experts had pointed out “striking similarities between the T1’s specifications and an already available, Chinese-made phone”.

CNN Business quoted another industry expert.

“Unless the Trump family secretly built out a secure, onshore or nearshore (fabrication) operation over years of work without anyone noticing, it’s simply not possible to deliver what they’re promising,” said Todd Weaver, CEO of Purism, one of the only known companies to actually manufacture a cell phone in the United States.

The dispute over the manufacture of this phone comes in the wake of President Trump one month ago heavily criticising Apple and CEO Tim Cook for moving production of iPhones from China to India. “I had a little problem with Tim Cook,” the president said at the time.

The Verge also noticed changes to some of the phone’s specs. “There are numerous errors on the page, from a processor section that doesn’t list a processor, RAM that’s described as storage, and the boast of a “5000mAh long life camera,” when it presumably means the battery.”

It noted that the triple camera array on the back looked similar to the iPhone Pro, except for the absence of a camera flash.

Originally published in Channel News Australia, 26 June 2025

The post Trump phone accused of being not ‘made in America’ first appeared on Chris Griffith - Technology Journalist, Australia.