Australia bans AI-made songs from its national music charts

Share
Australia bans AI-made songs from its national music charts

Australia's record industry drew a hard line on generative music this week, while new research shows state-linked hackers quietly industrializing their operations with open-weight models. Two very different answers to the same question: what happens when machines can do the creative — and criminal — work end to end.


Songs created mostly or entirely by AI are now barred from the ARIA charts, Australia's main music rankings, under eligibility rules that took effect this week. Releases must be "substantially human made": AI tools can still assist, but people have to write the song and perform the lead vocal and primary instruments. The trigger was DJ Josh Fawaz's AI-heavy cover of Madonna's "Like a Prayer," which topped the dance chart, peaked at number 2 overall and racked up more than 48 million Spotify streams before he added generative-AI credits amid a backlash. ARIA says the code exists to "promote the human nature of artistry," adding that it is "not interested in promoting or celebrating the success of AI-generated music that does not contain human artistry."

It's among the first major national charts to write AI eligibility into its rules, and the template other chart bodies will now be pressed to copy. The awkward part: a chart is meant to measure what people actually play — and in July, the most-played song on Australian radio was an AI-assisted track. Turning a popularity metric into a values statement is a defensible choice, but "substantially human made" will eventually need auditors, not just adjectives.


China-linked hacking groups have more than doubled their attack volume since putting open-weight models such as DeepSeek to work, according to Taiwanese researchers at TeamT5. The models handle reconnaissance, exploit writing and malware development — attractive because they're cheap to run and lightly guarded. Palo Alto Networks' Unit 42 recovered one session in which DeepSeek autonomously winnowed 25,209 exposed servers down to vulnerable targets, though it breached none: across more than 460 attempted targets, every confirmed intrusion was finished by hand. One crew even ran Claude Code while posing as an engineer conducting authorized tests to slip past its safeguards.

The reassuring read is that AI hasn't produced autonomous cyberattacks yet — human operators remain the bottleneck. The uncomfortable one is that the expensive part of hacking just got dramatically cheaper, and the gap between machine-scale scouting and human-scale breaking is precisely what the next model generation will shrink.

What to watch: whether Billboard, the UK's Official Charts and others follow ARIA's lead — and whether any chart body publishes an actual audit process behind "substantially human made."

Who should draw the line on how much AI a song can contain — the chart bodies counting the plays, or nobody at all? Tell us in the comments.

Read more

Kling 4.0 goes live with 30-second native video at 10-bit HDR

Kling 4.0 goes live with 30-second native video at 10-bit HDR

Kuaishou's Kling AI has spent two years becoming the video model people actually finish projects with. Its new flagship, Kling 4.0, went fully live on October 15 — and this round is less about prettier frames than about the two things that make a generated shot usable in a real edit: length and control. Kling 4.0 generates up to 30 seconds in a single native pass, outputs 10-bit HDR at 4K, 1080p and 21:9, and takes up to 15 reference assets in one task. That budget splits across ten images, fi

Brain organoids classified Braille at 83% — in a trio, not alone

Brain organoids classified Braille at 83% — in a trio, not alone

Two stories today: human brain tissue on a chip decoded a tactile signal, and the data-centre company behind some of the largest AI contracts in Europe turns out to be six months behind on its flagship site. Human forebrain organoids grown on a microelectrode array classified Braille letters with 61 percent accuracy on their own and 83 percent when the responses of three organoids were combined, in a peer-reviewed study from the University of Bristol. The team took event-based tactile readings

Suleyman says Microsoft's base model beat OpenAI's on a quarter compute

Suleyman says Microsoft's base model beat OpenAI's on a quarter compute

Microsoft finished pre-training its biggest in-house model yet — then used it to make a point about the company it bankrolls. Microsoft AI completed pre-training of MAI-Base-2, a mixture-of-experts foundation model with 80 billion active and 2.8 trillion total parameters, and says the recipe is 4.2–12.3x more pre-training FLOPs-efficient than DeepSeek-V4-Pro-Base. The run itself is the number worth keeping: 16,000 Nvidia GB200 GPUs for 38 days across August and September, at 21% average model

Amazon's new Bee devices listen all day — for $11.99 a month

Amazon's new Bee devices listen all day — for $11.99 a month

Amazon turned its always-listening Bee line into a two-device family on Thursday, and NVIDIA opened a 2.8-trillion-token Indic corpus recipe. The connecting thread: both are bets that the next AI money is in memory — yours, or your model's. Amazon put two new always-listening devices on sale, and both now lead to a subscription. Bee Frontier, the next-generation wristband that also clips to clothing, ships now at $74.99 in Onyx, Lavender and Sage — up from the $50 predecessor — with seven-day