TL;DR
- Chart Eligibility: ARIA can exclude or remove recordings when AI supplies all or the main creative performance.
- Human Work: AI-assisted tracks can qualify when people substantially create the recording and perform its lead vocal and primary instruments.
- Timing: The code first applies to the August 31 chart; no public ruling on Josh Fawaz’s track existed by August 26.
- Awards: A chart-ineligible title also loses ARIA Awards eligibility; separate 2026 award rules already require meaningful human authorship.
The Australian Recording Industry Association, the private trade body that runs the country’s principal music charts, will exclude recordings whose lead vocals, key instrumental performances or other primary creative elements were generated by AI. Recordings can still qualify when people made them substantially and performed the lead vocal and primary instruments, preserving a route for artists who use AI as a supporting tool.
ARIA’s boundary controls access to its own recognition systems: chart visibility and award eligibility. Copyright and streaming availability remain separate legal and platform questions.
The revised Charts Code is dated August 24, 2026. It first applies to the chart dated August 31, which is scheduled for publication on August 28. That timing makes Queensland producer Josh Fawaz’s AI-involved cover of Madonna’s Like a Prayer the immediate public case. ARIA had issued no title-specific ruling by August 26.
How the Human-Performance Test Works
ARIA’s two-part definition focuses on who or what supplied the recording’s defining performance. The AI-generated category includes recordings in which generative systems created all or the primary creative elements. ARIA names generated lead vocals, key instrumental performances and music produced wholly from a prompt as examples.
AI-assisted recordings sit on the other side of the line. A generative system may supply some expressive material, but people must still make the recording substantially and perform its lead vocal and primary instruments. The distinction makes the role of the tool more important than its mere presence: an AI effect inside a human performance is not treated the same way as a generated singer carrying the track.
Human performance is necessary, but it is not the only condition. The wider chart principles adopted by ARIA and other national chart bodies require authorised and lawful AI services, compliance with copyright and related-rights law, adherence to service terms, no reason to suspect chart manipulation and appropriate information that can follow the recording through the music supply chain.
Those conditions turn chart inclusion into more than a count of streams and sales. A recording can attract real listeners and still fail the private rules governing which consumption ARIA recognises in its rankings.
Why Josh Fawaz Became the Immediate Test
Artist Fawaz released an AI-involved version of Like a Prayer under his own name. By July, it had become the most-played track on Australian radio and appeared on several ARIA charts. Fawaz described AI as a tool in his process. Later Spotify credits acknowledged AI-generated vocals and drums, while an ARIA spokesperson said AI-generated lead vocals or key instrumentals would likely make a recording ineligible.
The recording remained a substantial chart performer immediately before the new code’s first application. On the ARIA chart dated August 24, Like a Prayer ranked fourth on the Australian Artist Singles chart, with 16 weeks listed. It also led the Australian Dance Singles chart that week.
That combination puts the track close to the rule’s clearest boundary. AI-generated lead vocals are one of ARIA’s explicit examples of a primary creative element. The association had not published a final ruling, and the pre-effect chart still included the song.
The code gives ARIA a process for resolving that uncertainty. Its Charts Department may ask a record supplier or data provider for information about a title. A supplier can seek a written eligibility ruling, and unresolved disputes can move from the Chart and Marketing Committee to the ARIA Board for a final decision. What the public documents do not yet show is how much evidence ARIA will demand when a human creator and a generative system both made material contributions.
What Ineligibility Changes
ARIA can refuse to survey an ineligible title, exclude it prospectively or remove it from past charts. The code also allows the association to adjust historical positions and withhold or revoke chart-related recognition, including No. 1 awards and sales accreditations, when the underlying recording fails the rules.
Annual ARIA Awards use a separate entry and judging system. The 2026 Awards criteria say only humans may be nominated or win, submissions with no human authorship are ineligible and AI cannot be the sole or central element of the work. A human must be the primary creator and make a substantial, meaningful contribution. The revised chart code adds another connection: a title that fails chart eligibility is also ineligible for any ARIA Awards.
Removal from a weekly chart changes a published ranking. Loss of a chart award or accreditation withdraws recognition attached to chart performance. Annual award eligibility determines whether a recording or person can enter, be nominated or win under that year’s criteria.
ARIA administers those charts and awards, not streaming catalogs. A track can remain available through a music service while failing ARIA’s eligibility test, because the code’s remedies change ARIA standings and recognition rather than a platform’s catalog. Platforms set their own rules for distribution, identity and recommendations, just as copyright questions are decided under law rather than by ARIA’s chart committee.
Why the Boundary Will Be Hard to Police
The rule’s poles are easy to describe. A song generated wholly from a prompt falls on one side; a recording sung and played by people with limited AI assistance can fall on the other. Mixed production workflows make the middle harder.
Automated tools have shaped recordings for years. Music professor Samantha Bennett noted that automated mixing and mastering plug-ins were already established. Fawaz’s recording presents a more consequential use: a generated lead voice and generated drums inside a release credited to a human producer. ARIA CEO Annabelle Herd defended the carve-out in similarly practical terms, saying artists already use AI tools and charts should “keep room for that”.
The disagreement is therefore not simply whether a computer touched the recording. It is whether the human contribution still supplies the performance that listeners and industry recognition attach to the named artist. ARIA’s lead-vocal and primary-instrument examples make that judgment concrete, but suppliers still have to document a workflow that listeners cannot necessarily hear from the finished track.
A separate US example shows why attribution can become as important as generation. The fictional AI country act Breaking Rust reached No. 1 on Billboard’s Country Digital Song Sales chart with Walk My Walk in November 2025. Associated Press reporting found that the vocal approach drew comparisons to country singer Blanco Brown, who said he had not known about the track. Breaking Rust did not prompt ARIA’s amendment, but the case illustrates what a chart can reward when the performed identity and the credited creator diverge.
The first chart under ARIA’s amended code is due to be published on August 28 for the week dated August 31. That chart, and any ARIA ruling on Fawaz’s recording, will show how the association applies its written human-performance criteria to a mixed human-and-AI release.
