- Develop a national sovereign
AI strategy: The ACS is calling for a "long-term
vision for sovereign AI development". However, I suggest this should be looking for short term measures. If Australia looks to the long term the strategy will be out of date before it can be implemented. As I suggested for the ACT Government last week, we can look to countries such as Singapore, for mature AI strategies to emulate.
- Develop an innovation strategy: The Australian Government is to release a "Strategic Examination of Research and Development", this year. Aligning workforce readiness with innovation by encouraging entrepreneurship is my day job. Government can do more to support startups and scaleups. One area is for defence, where the nation needs new capabilities quickly. Cancelling a few failed big ticket defence projects reliant on overseas suppliers could provide thousands of billions of dollars for local innovation.
- Government co-investment in scaleups: Not so sure about this recommendation, as it sounds dangerously like government trying to pick winners. Apart from strategically important areas, such as defence, energy security and availability of vaccines, government should avoid direct investment in scaleups, as they are really, really bad at it.
- Greater incentives for R&D in AI and tech adoption by businesses: ACS point out there is a skills gap in the use of AI by business. The obvious solution, I suggest, is training. Many of the computer project students I teach are planning to work in AI. To them it is not new and exotic, just a tool they are very familiar with. We need vocational education programs which similarly give business people a deep understanding of AI.
- Executives take a digital skills health check: Rather than trying to get C-suite leaders up to scratch on digital capabilities, I suggest political parties, government agencies and company boards need to hire more technocrats. We need some digital professionals running the nation and corporations, not just lawyers who learned a bit about computers.
- Promote entry-level pathways for cybersecurity professionals: ACS point out that we simply can't get enough computer professionals to fill cyber security positions. What we can do is take people qualified in allied fields and train them up. This can be done online with vocational education techniques, rather than in university lecture theatres.
- Implement an ‘earn while you learn’ scheme: ACS proposes business and government sharing the cost of worker retraining. Unfortunately what has tended to happen is as federal government introduces subsidies for university and vocational education, business and state governments have stopped funding. We have the technology to do the retraining (I spent a decade learning how this works). The problem is to make a watertight agreement on funding which business and states will abide by. Assuming Jobs and Skills Australia release their national skills taxonomy promptly, this can be used (if not someone else can). This can be aligned with the UK based Skills Framework for the Information Age (SFIA), used by ACS. The match will not be perfect. When ACS commissioned me to design a course in Green Computing (way back in 2008), there were no relevant skills in SFIA. I sent SFIA HQ a copy of my course and they added green skills, but that took considerable time.
- A national commitment to alternative tech pathways: ACS argues for greater trust in vocational qualifications by business. Another approach to this, I suggest, is the use of the Vocational Degrees, added to the Australian Qualifications Framework in February 2025. With this entry level staff can obtain a certificate and then go on to a degree at the same vocational institution. This should give business more comfort as to the depth of training provided.
Friday, August 1, 2025
Developing a national sovereign AI strategy
For the first time in a long time this week I missed the launch of the annual Australia’s Digital Pulse, by the Australian Computer Society at the National Press Club (earlier in the day I twisted my knee running for a bus). The report has 10 recommendations. My comments on some:
No comments:
Post a Comment