HUGH ‘POP’ SELBY shares a ‘leaked’ look to a future world where artificial intelligence starts gnawing its way through education and the legal profession.

The following document, written in April, was intercepted by a precocious 12-year-old who sees every computer security system as a challenge. 

She wanted to boast about her success to her grandparents. She did this first with her nana who was in the kitchen, hand beating a sponge cake mix. Then she went out to share the document with her pop who was slowly digging up and dividing daffodil bulbs.

It was he who emailed it to the CityNews editor with a cover note. 

Memo

To: Federal and Territory education planners at primary, secondary and tertiary levels;

And: Federal and Territory ‘cost cutters’ in Education and Justice Departments.

Importance: High

Readability: Easy

Classification: Restricted

Date: World Tech Day, April 31 2024

Context

Infrastructure cost blow outs, such as AUKUS and the light rail, and what will have to be spent in the next decade to make up for the growing nationwide housing shortfall, entail that savings must be made in other areas.

Fortunately, there are significant cost savings to be made now and in the next five years in our education system and in the delivery of “justice” services.

These savings flow from the rapid, dramatic advances in AI capacity. The rate at which AI is challenging conventional approaches to education and legal services has been so rapid this past year that most people working in these areas have not noticed how much their world has changed already, and will change again before this time next year.

Although this memo discusses only education and legal services, there are other areas of government expenditure where, no doubt, subject matter experts can point to areas where significant cost savings can be made.

Education

For teaching from age six onwards we can now do away with small classes. Just give every child headphones and a tablet loaded with high-quality teaching materials with built-in assessment.

 All the students at the same level can be seated together while working individually on the pre-loaded interactive materials. Of course, many of them may work from home, especially if one or both parents work from home. Those covid isolation days were just a taste of what is possible, especially as AI offers so much better teaching materials than was possible even a year ago.

A student progresses to the next topic when they master the present one. A room assistant (much cheaper to employ than a qualified teacher) monitors all real-time assessments to pick out the student/s who are falling behind. They are taken out for short sessions, small group teaching by a human teacher.

The results for those students after “remedial” work can then be plotted for each teacher. Those teachers who fail to bring their students up to the required standard within an appropriate period can be made surplus to requirements.

Because of the increasing rate of technological change it is recommended that all new teachers be placed on a three-year probation – so that those unable to adapt quickly to changing teaching demands can be got rid of quickly and cheaply.

The speed with which a child acquires learning can also be tracked so that, for example, if they are well ahead of their age cohort in one subject, they can have small group human teaching in that subject with students at their level (whether the same age, older or younger) to better motivate them to excel.

This can be done from early primary to the end of high school. We can do away with selective schools save for those children/adolescents who excel in sport. They need to be brought together in one place so that their coaching regimes can be well managed.

AI online, interactive teaching brings a more individual focus to the student, frees up teachers to give attention where it is most needed, creates more flexibility in subject offerings, helps achieve good building space utilisation, ensures better real time assessment of students’ progress, and saves time and money.

Bottom line: fewer teachers, fewer classrooms, larger classes, less cost.

Law schools

Law schools will shrink in number, as will the need for lecture theatres. Proximity to where students live and work is unnecessary. Subject specialists from around our nation can get together, and with the help of AI and audio visual experts, convert dull legal teaching into engaging, problem recognition and problem solving online learning materials.

No matter where the student is enrolled, they can access top-quality materials. As in schools the assessment will be immediate, progressive, and come back with AI-generated comments designed to enhance and embed the law student’s knowledge. Semester and/or annual examinations will be no more. Essays too are a thing of the past, making plagiarism a non event.

Some reactionaries will bemoan the loss of writing skills. The answer, of course, is that with A1, and editing programs, learning such a skill is unnecessary.

More use will be made of current and recently retired legal practitioners who for little money (there being a large oversupply of underworked lawyers) will value-add to the learning experience by online sessions that make use of their practical skills to show students how to customise template contracts, wills, property leases etcetera for different client needs. Students can access these in real time when they can contribute, or out of hours when they are passive learners.

Freed from boring, mass teaching and assessment tasks, law academics can be committed to useful research and running short-course intensives in subspecialties where there is a focus on student research and high-level discussion of policy issues in the law. 

Bottom line: much less face-to-face teaching, more time for academic research, students acquire problem solving skills throughout each subject, abolition of exams and essays, closure of some law schools, savings in academic salaries and building construction and maintenance costs, excess number of law graduates will keep entry to work salaries low and encourage unpaid internships.

Courts and tribunals

Court complexes, judicial officers, court staff, jurors add up to a large impost on the public purse. AI means that this can be trimmed, as the following examples show.

“Pay by direct debit or EFT” infringement notices (for example, parking, speeding, littering) are a tidy little earner for government revenue. The approach can be extended further into all those matters where defendants are required to attend court.

Around 90 per cent of all criminal matters finish on a plea of guilty. Many of those who punt on a possibly successful defence could be persuaded to plead guilty. Here’s how. 

For many people a realistic approach to the most likely, but unwanted outcome comes at the last minute when reality cannot be avoided. In litigation, this is on the steps of the courthouse. Hence, in the foyer of the Magistrates Court install an array of AI machines. These are preloaded with the latest data on penalties and how various factors influence the penalty outcome. The defendant completes a questionnaire at the machine which then advises the likely penalty.

If there is an assessed likelihood of actual jail time then the defendant must go into court in front of a magistrate.

But for any lesser penalties such as a fine, even supervised release on conditions, the punter can accept that result and proceed to payment, or front up at Community Corrections to enter into the supervision arrangement. 

Alternatively, the punter can take their chances, reject the machine result and go into court. However, if the result is a penalty that is the same as, or greater than the machine advice, then it is user pays for the human judge’s involvement; that is, the punter receives the bigger penalty and pays court costs.

Able-bodied offenders who can’t pay, or won’t pay, should be pressed into manual labour where they cut out their financial obligation at a daily rate that is one fifth of the current unemployment benefit. Our parks and gardens, road sides, culverts, and bus interchanges will all benefit.

AI will calculate the defendant’s odds of success and state what they are. The plea of guilty rate being around 90 per cent, every 1 per cent shift from not guilty to guilty plea is a 10 per cent reduction in hearings.

Not only will this mean that fewer courtrooms, judicial officers and jurors are needed, it will also cut down on the Legal Aid workload and hence its budget needs.

The Federal budget “legal services” bottom line can also benefit from AI. To reduce the hearings for Family Court financial settlements all that is required is to load up the financial data from both parties and the AI machine will spit out a likely distribution that takes account of factors such as length of relationship, nature of contribution (financial and otherwise) to that relationship and special needs. This approach gets rid of the expensive pre-trial mediations.

As in the territory courts, it will be user pays: if a party or both parties choose to take their chances in front of a human and the result is equal to, or not as good as what AI predicted, then the punter/s pay a hefty fee.

Running an appeal in both civil and criminal matters, be they territory or federal based, will be much simpler. The party thinking about appealing can feed into the AI machine the entire hearing transcript, plus judge’s directions and, in judge alone trials, their decision. AI will spit out possible grounds of appeal, the linking transcript/judgment references, and an odds to win prediction.

The same approach can be taken to cases before Tribunals: for example, end of tenancy disputes, repair disputes, consumer law, disputes with ACT Revenue or the ATO, immigration issues, etcetera.

Bottom line: fewer judicial officers and tribunal members, fewer hearing rooms, shorter waiting times to determine disputes, increased revenue from fines and fees.

A note to readers: pop’s cover note to the Editor was short. “I am glad to be a mere spectator in the race to replace human-to-human interaction.”

Former barrister Hugh Selby is the CityNews legal columnist. His free podcasts on “Witness Essentials” and “Advocacy in court: preparation and performance” can be heard on the best known podcast sites.