Personal
The person's own values, memories, choices and preferred models.
Personal AI.
On your terms.
Personal AI could help you understand information, reflect on experience and make choices. Mutual Futures connects that capability with GAJRA’s question of alignment: does the system help you act on what matters to you, under permissions you can understand?
Begin with a person’s purpose
Aura of Intelligence is a developing design for personal AI. The aim is to help a person organise knowledge, reflect on experiences and choose what to share. A useful starting task might be preparing for a new role, finding a source or comparing a decision with personal priorities.
The project’s proposed cognitive architecture remains research and development. Its technical design is documented in the Aura project; this site focuses on the person’s choices and the work a system would need to perform.
Sources for this section
P02 · Aura of Intelligence D13 · UNGA81: Joyful Responsible Abundance
Use computing at the scale the task needs
Personal devices, shared local facilities, regional infrastructure and cloud services can each have a role. Choose around the task, operating cost, energy, access and the information involved. Shared computing could support enterprise design and learning without making every person dependent on one provider.
Sources for this section
D03 · A Fair Go for the AI Age D13 · UNGA81: Joyful Responsible Abundance
Connecting personal AI with legal information
A future Aura system could help people find and understand legal sources. The Australian Law 2012 project explains the original approach, and a separate Australian Legal Engine repository describes tools for finding and tracing legal provisions. Integration with the Aura app remains unfinished.
Sources for this section
P08 · Australian Law 2012: Luke's Relevance P09 · Australian Legal Engine
Choose how information may be used
Using information to answer a question is different from using it to train an AI model. Keeping it on an approved local system is different from sharing it with an external service. The case study below explains why those choices matter. The builder records preferences; it does not enforce them.
Professional
The qualifications and relevant information a role genuinely requires.
Collective
Material whose custodians or communities determine the relevant authority and permissions.
Make responsibilities clear
Before a person, AI system or robot carries out a task, the people involved should understand its purpose, who authorised it, what skills and information it needs, and who checks the result. There also needs to be a way to stop, correct mistakes and recover.
Why sharing choices matter: the CarCover example
The OAIC published a fictionalised case based on a reported data breach. An employee uploaded a customer’s hardship application, including health and family details, to a public AI service despite a workplace policy against it. The example shows why purpose, sharing and training choices need to be understood before use.
Why the tool starts this way
The draft starts with local use only, no training permission, no external sharing permission and a limited purpose and period. These are this site’s cautious starting choices in response to the case, not settings used by the company or a guarantee of compliance. The type of information remains undecided until you choose it. Saving this draft does not configure any AI service.
Draft your preferences
My permissions, in plain language
Choose how your information may be used without entering private records. Download a draft to discuss with a provider or workplace. It does not change their systems or enforce your choices.