Ben Reid

Reposting an article by Alan Morrison on #SystemsThinking. Systems Thinking promotes a holistic view of connected and responsive operations ("living systems"), which makes it ideal for today's increasingly fragmented #EnterpriseIT estate and the #complexdistributedsystems that have emerged. When I did my foundational research before launching EnterpriseWeb I read up on #complexitytheory, #complexadaptivesystems, #cybernetics, #OODA loops, #controltheory, #coordinationtheory, etc. As Alan suggests, #graph technology is key to implementing dynamical and contextual systems. However, #GraphDBs are poor candidates. "Living Systems" depend on connections to the real-world and the ability to evolve with it. GraphDBs aren't really made for connecting and transforming real-time data between services, systems, databases and devices (classic EAI, ETL, orchestration). GraphDBs also don't do well with temporal history and they don't model behavior or software contracts. In separate recent threads, I objectively described GraphDB limitations for such use cases (https://lnkd.in/eJvqFHHZ) and the solution approach we took (https://lnkd.in/eY3mTxgP).

3

Interesting question posed by Alex Roberts in Australian Gov's Innovation Month's national discussion on Reimagining Public Services with Generative AI. Alex asks, "What other core assumptions will be challenged or shown to be false in a world of generative AI? Which core assumptions are no longer fit for purpose?" Two years ago we could safely assume that: * Public consultation submissions were written by the people submitting them (or that you could tell if they were not) * Complex information could not be easily distilled and tailored to different audiences * Images and video could generally be trusted as being authentic * A person’s voice would be hard to replicate * Developing AI solutions required extensive training and skill. These assumptions are no longer valid (or at least not as much as they were). OpenAI has already encouraged the phasing out of voice-based authentication as a security measure for sensitive information, given its experience with their Voice Engine." Larry Marshall Tony Surtees Sir Geoff Mulgan Tom Lounibos Jim Varghese, AM Stephen Cheung #GovInnovationMonth #Innovation #PublicSector #GetInvolved #DigitalProfession

15

1 Comment

Public Sector AI lessons from New South Wales (NSW), Australia : 🥇 First in Australia to implement a whole-of-government AI Ethics Policy and AI Assurance Framework 📊 Released its AI Strategy in 2020 🌟 Contributed to Australia's impressive #5 ranking in the OECD Digital Government Index 💡 Fun fact: There's no universally accepted definition of AI! NSW gov defines it as "intelligent technology, programs, and advanced computing algorithms that augment decision-making." 🤔 What are your thoughts on AI governance? How is your region approaching AI regulation? #ArtificialIntelligence #AIGovernance #DigitalTransformation #NewSouthWales #TechPolicy #AIinpublicsector

10

Digital Twins are not a new concept, but for years, the conversation has centred around definitions. While the specifics may vary, most agree that a Digital Twin involves a two-way connection between a physical asset and its digital counterpart, which represents its attributes and behaviours. However, if you're looking to deepen your understanding of what makes up a Digital Twin, I highly recommend this excellent short video from Daniel Block at the Connected Places Catapult Digital Twin Hub. It breaks down the different components that come together to create a Digital Twin in a clear and accessible way. The reality is, few organisations start with a fully realised Digital Twin. Most are on a journey—adding additional components over time as their needs and capabilities evolve. Whether you're just starting or refining your approach, understanding these building blocks is crucial to unlocking the full potential of Digital Twins. #DigitalTwin #DataSharing #GeminiReport #Collaboration #Innovation Digital Twin Hub at CPC Dr Alison Vincent, CBE FREng Dr Wolfgang S. Raj Mack David Flynn Ryan Hood Erika Lewis Jonathan Eyre Sarah Hayes Ben Ramsden Alan Welby Simon Evans Fergus Cloughley John Ahmet Erkoyuncu Holger Kessler Fionn McGorry CIPP/E George Economides Tim Hare Alex Gluhak Dr. Stefano Tedeschi Sophie Peachey Andrew Collinge Theo Blackwell MBE https://lnkd.in/eT7R7Z_k

33

Why Australia Needs Its Own AI Large Language Model I loved Anton van den Hengel's article in the The Australian Financial Review , advocating for #Australia to develop its own AI large language model. Here are some key points that resonated with me: National Sovereignty and Security: Developing an AI large language model locally ensures that Australia maintains control over its data and intellectual property, mitigating the risks associated with relying on foreign AI technologies. Economic Competitiveness: By creating its own AI infrastructure, Australia can drive innovation across various sectors, including healthcare, finance, and education, positioning itself as a global leader in AI technology. Tailored Solutions: A locally developed AI model can be fine-tuned to address specific Australian contexts and linguistic nuances, enhancing its relevance and effectiveness for Australian businesses and government agencies. Job Creation and Skills Development: Investing in AI research and development will create high-tech jobs and foster a skilled workforce, crucial for sustaining long-term economic growth and competitiveness. Personal Insights: As someone deeply involved in the tech sector, I fully support the push for Australia to develop its own AI capabilities. Not only does this align with national interests, but it also opens up a wealth of opportunities for innovation and growth. Embracing this challenge will ensure that Australia remains at the forefront of technological advancements, driving both economic and social benefits. What do you think about the idea of Australia developing its own AI large language model? How do you see it impacting our economy and tech landscape? #Innovation #AI #Technology #Australia #EconomicGrowth #FutureOfWork #tech4humanity Feel free to reach out to discuss past, present, or future activities. Link to article in the comments

44

12 Comments

The art and science of summarization is critical to human knowledge acquisition. Generative AI developers thought that summarization can be done without understanding the context or semantics by inductive means using LLM. An article “Australian government trial finds AI is much worse than humans at summarizing” states AI summaries were judged significantly weaker across all five metrics used by the evaluators, including coherency/consistency, length, and focus. It failed to capture "complex context". That finding does not come as a surprise since other researchers came to similar conclusions. Human has been summarizing text since there is natural language communication. We call it hearsay, regurgitating the gist or paraphrasing. In general hearsay is not admissible in court for good reasons, but direct quotation extracted from actual text is admissible. The reason that summarization made with Generative AI is inferior is not difficult to understand. It uses an inductive technique with input document and pre-trained model to generate words, in hope to produce similar context and semantics with a smaller number of words than the original document. Unless the same input document or something highly resemble to inputted document were “pre-trained” into the model, the result is weak, especially in the area as identified by Australian Government’s trial. The ideal summarization approach should be one that is repeatable and comprehensive in nature while supported by annotation through references to the actual words that depict the neighborhood semantics. Implementation logic is typically derived from deductive reasoning supported by propositional logic. ELAINE is an example of such implementation. Besides the science behind good summarization, we also need a good metaphor that human users can get to the information quickly such as knowledge map, high level abstraction and logical relations. https://lnkd.in/gik4QGy4

65

24 Comments

QUANTUM? TELL HIM HE'S DREAMING Quantum might make 20,000 jobs and $6B in revenue by 2045 according to CSIRO. Excellent. Australia should get in on the action. If it pans out, life will be grand. Meanwhile, TODAY!, our software industry already dwarfs those numbers and for a fraction of the eyewateringly huge bag of cash we're already throwing at #quantum we can easily turbo-boost the #software and #SaaS sectors. 2,000 new software ventures started this year alone and this trend will continue for the foreseeable future. So that will be hundreds of thousands of jobs and billions in revenue long before 2045. Perhaps we should spend a little less time looking misty-eyed at a possible future, and instead take a long hard look at the reality of today. #MrSaaSSays - How's the serenity? https://lnkd.in/ePkbdYeG

7

2 Comments

“...you’re not just talking circular economy, you're actually talking MOLECULAR CIRCULARITY …. not just about thinking about Recycling & Isolation, it's coupling that, and align that, with Manufacturing & Remanufacturing. The fabulous example is our plastic filaments #MICROfactorie…  plastic that came from the I.T. sector (which is typically considered really difficult to recycle)… we can indeed convert that plastic into… PLASTIC FILAMENTS in our #MICROfactorie - and of course you know we know that plastic filaments then become a FEEDSTOCK for 3D printing!" Engineers Australia's Create Magazine's exclusive sit-down interview with Professor Veena Sahajwalla. #australianinnovation #circulareconomy #plastics #localisation #decentralisation #remanufacturing Manufacturing Excellence Forum Cleantech Network

3

2 Comments

More independent, scientific testing showing that good quality #facialrecognition & #liveness software does not show significant bias. The testing agency was BixeLab, the software vendor was Daon & the test set was ethnic groups in New Zealand. Biometric Update explains: “The report from BixeLab on accuracy differentials in New Zealand’s biometric identity verification platform for public service access has been published, showing the system is not biased, but does have slightly higher false positive rates for Māori and Pasifika people. The IdentityX cloud platform and liveness detection software were found in the “Level B Evaluation Assurance — Identity Check Bias Evaluation” to have “no significant bias” in liveness or matching performance “among the tested ethnic groups.” Those groups include Māori, Pasifika and “other” New Zealanders.

11

6 Comments

CAN WE? YOU BETCHA... ...If we don't starve in the meantime. "I see Australia’s strength in high-quality tertiary education to be our superpower." says Stephen Bartlett Damn right. We have the education system and we're not a poor nation. All we need now then is some proper commercialisation support. But, while I'm a huge fan of the quantum work, it feels like we're trying to land a whale, and who knows, perhaps we will. But we need to get fed while we're waiting and our thousands of burgeoning software businesses could do with a little help too, and they'll cost way less to support than anything to do with quantum. Follow SOFTWARE Australia today and help us build a software industry to take on the world. https://lnkd.in/d_3r844i

2

1 Comment

Super exciting to see such strong momentum in the IIoT space! 🌟 Our team at Boodskap IoT shares your enthusiasm—IIoT is indeed at a critical juncture, and the potential to leverage the vast amounts of data being generated is immense. As we continue to innovate and align our platform with industry needs, it's inspiring to see other leaders pushing the boundaries as well. Looking forward to connecting and exchanging ideas with fellow pioneers in this space. The future of IIoT is bright, and together, we can make it a reality! 🚀 #IIoT #Innovation #DigitalTransformation

16

Pro Tip: "Partner with people smarter than yourself". In this fantastic post Patrick Mandic perfectly illustrates that I've taken my own advice ;)

7

https://lnkd.in/eEyGFEXJ   Technology company DAS has released its first public, detailed region-by-region growth snapshot for winter crop production. Utilising satellite technology and machine learning, DAS provides public measurement of Net Primary Productivity (NPP), a biophysical estimate of plant growth rate that corresponds with crop yield. This new data complements broader state estimates previously released by government agency ABARES. According to DAS, the figures form part of a broader suite of Grain Intelligence reporting metrics. These metrics offer detailed estimates for crop area, yield, production, hay area, and harvest progress, with flexibility in defining reporting regions. The models are derived from over 109,000 ground truth locations, including more than 13,000 locations from the year 2024. Manufacturing – Australia – August 2024 - https://lnkd.in/eqUFHE6k DAS (Digital Agriculture Services) #satellite #Expertfarming #Sustainable #agriculture

7

1 Comment

Why we need a catalogue of tools and metrics for trustworthy AI There are tools and metrics out there that help AI actors to build and deploy AI systems that are trustworthy. However, these tools and metrics are often hard to find and absent from the ongoing AI policy discussions. This catalogue makes it easier to find tools and metrics by providing a one-stop-shop for helpful approaches, mechanisms and practices for trustworthy AI. https://lnkd.in/ec9DdaTf

5

Quantum Computing: Future Potential, AI: Here & Now. There’s been a lot of buzz around Quantum Compute of late. Especially with the Australian Govt recent somewhat controversial announcement of almost A$1bn in funding for the US-based quantum-startup PsiQuantum. The investment makes PsiQuantum the world's most funded independent quantum company. Whilst much of the technical qubit discussion is admittedly above my pay scale, there's a pressing reality we perhaps can't ignore: quantum computing, according to many experts, is still years away and with uncertain outcomes. Meanwhile, AI is here now, delivering real results that we shouldn’t fall behind on. A recent study by researchers from Microsoft and the Scalable Parallel Computing Lab at ETH Zurich (SPCL@ETH) provided a harsh reality check for those hyping the potential of quantum computers. They found that off-the-shelf GPUs outperform these machines in practical applications. Their paper demonstrated that, short of exceptional improvements in hardware & software, even future quantum systems are unlikely to achieve practical speeds in workloads like drug discovery, material sciences & weather forecasting. To highlight this, they pitted a hypothetical quantum system with 10,000 error-correcting qubits against a classical computer equipped with a single NVIDIA A100 GPU, and the GPU came out well ahead. Sovereign Large Language Model (LLM) AI capabilities are paramount for maintaining Australia’s competitiveness & technological autonomy. Whilst A$1BN is being spent on PsiQuantum, experts emphasise that Australia could establish a sovereign AI capability for around $100 million in compute costs. In a recent Senate inquiry into AI adoption, Professor Nicholas Davis from the University of Technology Sydney explained that Australia has three key options, being: - Accept US market-leading LLM’s with their transparency & data set limitations. - Use open-source models and accept a big performance gap. - Develop a sovereign LLM using Australian data & experts. Greens Senator David Shoebridge highlighted that a $100m investment is a surprisingly low price for such critical sovereign capacity, yet the Government's AI funding commitment in the last Budget was ~$40m, primarily for regulatory efforts over 5 years. Simon Bush, CEO of the AIIA, noted this as insufficient, comparing it to the UK’s £2bn, Singapore’s US$5bn, & Canada’s US$2.4bn investments. Bush warned that without significantly more funding and urgency, Australia risks falling well behind its global competitors. It's clear: Quantum Computing is on the horizon, but AI is here, demanding immediate attention & investment. The question is, will we seize the opportunity or watch it slip by? https://lnkd.in/gR2tUK-f https://lnkd.in/gn_znJPY #AI #QuantumComputing #SovereignAI #GreenSquareDC #NVIDIA

36

5 Comments

Embracing #AI: The Path to Transforming Public Services& Governance 𝐀𝐫𝐭𝐢𝐟𝐢𝐜𝐢𝐚𝐥 𝐈𝐧𝐭𝐞𝐥𝐥𝐢𝐠𝐞𝐧𝐜𝐞 (𝐀𝐈) 𝐑𝐞𝐯𝐨𝐥𝐮𝐭𝐢𝐨𝐧 𝐢𝐧 𝐊𝐧𝐨𝐰𝐥𝐞𝐝𝐠𝐞 𝐖𝐨𝐫𝐤: - Efficiency 𝐺𝑎𝑖𝑛𝑠: GPT-4 & other generative AI models enable knowledge workers to increase task completion by 12%, reduce task completion time by 25% & significantly boost productivity, especially for less-skilled workers. - 𝐵𝑢𝑠𝑖𝑛𝑒𝑠𝑠 𝐼𝑚𝑝𝑎𝑐𝑡: Companies leveraging AI tools are 40% more efficient & enjoy 70% higher customer & employee satisfaction rates compared to those that don't use AI. 𝐀𝐈 𝐈𝐧𝐯𝐞𝐬𝐭𝐦𝐞𝐧𝐭 𝐚𝐧𝐝 𝐀𝐝𝐨𝐩𝐭𝐢𝐨𝐧: - 𝐶𝑜𝑟𝑝𝑜𝑟𝑎𝑡𝑒 𝐼𝑛𝑣𝑒𝑠𝑡𝑚𝑒𝑛𝑡: Leading tech companies are planning over $250Bn investments in AI infrastructure, with total corporate AI investments since 2020 nearing $1 trillion. - 𝑊𝑖𝑑𝑒𝑠𝑝𝑟𝑒𝑎𝑑 𝑈𝑠𝑒: Within 9 months of its launch, ChatGPT was used by 80% of Fortune 500 companies. European spending on generative AI is projected to grow by 115% in 2024, & global AI investment is expected to reach $200 billion annually by 2025. - 𝑀𝑎𝑟𝑘𝑒𝑡 𝐺𝑟𝑜𝑤𝑡ℎ: The global AI market is anticipated to exceed $1 trillion by 2028. 𝐏𝐨𝐭𝐞𝐧𝐭𝐢𝐚𝐥 𝐢𝐧 𝐭𝐡𝐞 𝐏𝐮𝐛𝐥𝐢𝐜 𝐒𝐞𝐜𝐭𝐨𝐫: - 𝑇𝑟𝑎𝑛𝑠𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑣𝑒 𝑃𝑜𝑡𝑒𝑛𝑡𝑖𝑎𝑙: AI can vastly improve public services, enhancing efficiency, reducing backlogs, & improving citizen-government relationships. The UK government estimates that up to a third of public-sector tasks could be improved with AI, potentially yielding £40Bn annually in productivity gains. 𝐒𝐭𝐫𝐚𝐭𝐞𝐠𝐢𝐜 𝐑𝐞𝐜𝐨𝐦𝐦𝐞𝐧𝐝𝐚𝐭𝐢𝐨𝐧𝐬: 1. 𝐴𝐼 𝑀𝑖𝑠𝑠𝑖𝑜𝑛 𝐶𝑜𝑛𝑡𝑟𝑜𝑙: - Establish a centralised AI team in Number 10, led by an AI Mission CEO, to coordinate AI efforts across government departments. 2. 𝐼𝑚𝑚𝑒𝑑𝑖𝑎𝑡𝑒 𝐺𝑜𝑣𝑒𝑟𝑛𝑚𝑒𝑛𝑡 𝐴𝑐𝑡𝑖𝑜𝑛𝑠: - Assign the Chief Secretary to the Treasury to oversee digital transformation, data, & AI. - Reform the National Audit Office's evaluation approach to encourage risk-taking. - Update civil-service career frameworks to attract AI talent & streamline recruitment. - Designate “AI exemplar” departments to lead by example in AI adoption. - Enforce a “Bezos mandate” to ensure all departments provide secure data access. 3. 𝑇𝑒𝑐ℎ𝑛𝑜𝑙𝑜𝑔𝑖𝑐𝑎𝑙 𝐹𝑜𝑢𝑛𝑑𝑎𝑡𝑖𝑜𝑛𝑠: - Prioritise data interoperability & invest in sovereign AI capabilities. - Secure necessary computing power and collaborate with private-sector partners. - Compete for top AI talent by offering competitive salaries & streamlined hiring. 4. 𝐷𝑒𝑝𝑙𝑜𝑦 𝐸𝑥𝑖𝑠𝑡𝑖𝑛𝑔 𝐴𝐼 𝑇𝑜𝑜𝑙𝑠: - Use AI to alleviate backlogs & waiting lists in public services, like the NHS and education, to save costs and improve efficiency. 𝐋𝐨𝐧𝐠-𝐓𝐞𝐫𝐦 𝐕𝐢𝐬𝐢𝐨𝐧: - AI-Enabled Govt: A government where every citizen has a digital assistant to manage interactions with public services, ensuring a faster, higher-quality, & more cost-effective state.

1