Originally published at: Can AI unlock human productivity? – InvestEngine Insights
AI is emerging as the transformative force that will redefine human productivity. From improving workforce efficiency to accelerating knowledge creation, AI promises an era of economic and scientific progress. This essay explores the optimistic future of AI’s impact on productivity, supported by recent studies and case studies from a variety of sectors, illustrating how AI has the potential to revolutionise not just industry but also society.
The rise of domain-specific foundation models
I believe it’s unlikely that AI’s future will be dominated by a single, all-powerful model but by an ecosystem of domain-specific foundation models. These specialised AI systems, trained on data from specific fields, are already enhancing productivity in areas like climate science and drug discovery. For example, Microsoft’s Aurora model is revolutionising weather forecasting through comprehensive simulations that utilise 3D spatial data.[1] The model improves accuracy in predicting rare and extreme weather events, which has a direct impact on industries reliant on precise climate data, such as agriculture and logistics.
Similarly, Chai-1, a foundation model designed for drug discovery, is capable of predicting molecular structures, enabling researchers to accelerate the development of new medicines.[2] With drug discovery being notoriously time-consuming and expensive—requiring billions of dollars and years of trials—AI systems like Chai-1 can significantly reduce costs and timelines. By improving success rates in early-stage clinical trials, Chai-1 demonstrates how AI can optimise research and development, contributing to productivity gains in the pharmaceutical industry.
“Chai-1’s ability to accurately predict molecular structures and interactions opens new frontiers in drug discovery. It not only improves accuracy but also allows us to analyze complex molecules that were previously beyond the reach of traditional AI models” – Dr. Emily Rhodes, CTO of Chai Discovery
These domain-specific models are just the beginning of AI’s potential to drive productivity in specialised industries. I expect their ability to reduce complexity and increase precision makes them indispensable tools in the future of scientific research and industrial efficiency.
AI’s immediate impact on workforce efficiency
AI has already shown its ability to enhance worker productivity, especially those with less technical expertise. A study by Boston Consulting Group revealed that non-specialist employees, when augmented by AI, achieved performance levels comparable to those of data scientists.[3] For instance, in coding tasks where non-specialists typically scored 37%, AI assistance raised their scores to 86%.[4] This remarkable improvement highlights AI’s potential to elevate workforce productivity, enabling companies to achieve more with less.
However, the study also revealed that AI does not inherently improve workers’ baseline skills. Without AI assistance, the augmented workers reverted to their original performance levels. This finding suggests that while AI can significantly boost productivity in the short term, long-term gains will require a strategic focus on developing critical thinking and problem-solving skills. The ability to break down complex tasks using Aristotelian “first principles thinking”, will be essential to prevent workers from becoming overly dependent on AI.
In this context, AI serves as a tool that enhances the workforce by enabling non-experts to perform at much higher levels. This will be especially crucial in industries facing skilled labour shortages or where rapid upskilling is not feasible.
AI and the expansion of knowledge creation
AI’s contributions to knowledge creation are perhaps its most exciting potential. Historically, scientific advancements have often been limited by the tools and methodologies available to researchers. Today, AI is providing the tools to accelerate discoveries in fields as diverse as quantum chemistry, biology and physics.
AI’s ability to generate new ideas and simulate complex phenomena is illustrated by DeepMind’s Alphaproteo which designs novel proteins.[5] Similarly, NVIDIA’s Megatron-BERT model is being used in drug discovery to simulate molecular interactions and predict how compounds behave in biological systems.[6] Both tools are not merely augmenting human capabilities but are bridging gaps between disciplines, enabling researchers to cross-pollinate ideas more effectively.
As the complexity of scientific knowledge increases, AI will evolve from a tool to a critical partner in discovery. More than just augmenting human capabilities, AI will enable innovation across disciplines, empowering scientists to tackle the most complex challenges of our time, from climate change to global health crises. AI may even play a role in helping us unlock deeper questions about the nature of consciousness itself, one of humanity’s most profound mysteries.
Economic and business productivity gains
Beyond its impact on research and science, AI is poised to deliver substantial economic gains. Bain & Co predicts that integrating generative AI into business operations could boost EBITDA by as much as 20% within 18 to 36 months.[7] This rapid payback period contrasts with earlier technological revolutions, such as the adoption of electricity or PCs, where productivity gains took decades to materialise.[8]
In my view, the speed at which AI is delivering ROI in specific sectors is impressive. For example, in software development, AI-driven tools like GitHub Copilot have increased developer efficiency by 26%.[9] Similarly, Amazon has demonstrated that developers using Amazon CodeWhisperer, which serves a similar function to Microsoft Copilot, completed tasks 57% faster and were 27% more likely to complete them successfully compared to those who did not use it.[10]
Furthermore, AI’s ability to automate complex tasks and chain multiple skills together—what experts refer to as “agentic systems”—will enable businesses to tackle long-term projects with greater efficiency. By automating workflows that require adaptability and problem-solving, AI will fundamentally reshape how companies operate, leading to sustained productivity growth.
Technological infrastructure and the future of AI
As AI continues to evolve in complexity, concerns about the availability of computing power have emerged. Companies like Google, which now reportedly spends more on computational infrastructure than on human capital, illustrate this shift in focus.[11] However, I believe the real challenge lies not in a lack of resources but in the intelligent deployment of infrastructure.
The future of AI productivity hinges on leveraging advanced infrastructure, such as specialised hardware (GPUs, TPUs) and optimised data centres, to meet the increasing demands of AI models without sacrificing efficiency. Tech giants like Microsoft are also investing heavily in next-generation data centres that focus on sustainability and efficiency.[12] As AI becomes more integral to business success, the companies that build atop of this infrastructure will lead the next wave of innovation.
Conclusion: AI as the driver of long-term productivity and innovation
AI is on the cusp of transforming productivity in profound ways. Its ability to enhance workforce efficiency, accelerate knowledge creation and spur economic growth is undeniable. Domain-specific models, like Microsoft’s Aurora and Chai-1, demonstrate how tailored AI systems can revolutionise specialised industries, while AI’s impact on workforce productivity shows how even non-specialists can perform at advanced levels. In science, AI is poised to break through the limits of human cognition, fostering interdisciplinary collaboration and accelerating discovery in fields previously constrained by human resources and knowledge.
The economic benefits of AI are already beginning to materialise. As AI systems become more integrated into business processes, the productivity gains will grow. Moreover, concerns about a potential compute crunch are being alleviated by advancements in infrastructure, ensuring that AI’s computational demands will be met for the foreseeable future.
Therefore, AI represents a monumental opportunity for humanity to transcend its natural limits and elevate productivity to new heights. Whether in the lab, on the production floor, or in the office, AI will drive future growth. The key question now is how quickly businesses and societies can harness its full potential to tackle the challenges of tomorrow. Are we prepared for what lies ahead?
Important information
Capital at risk. The value of your portfolio with InvestEngine can go down as well as up and you may get back less than you invest. ETF costs also apply.
This communication is provided for general information only and should not be construed as advice. If in doubt you may wish to consult a professional adviser for guidance.
Tax treatment depends on personal circumstances and is subject to change, and past performance is not a reliable indicator of future returns.
[1] Microsoft, “Introducing Aurora: The first large-scale foundation model of the atmosphere”, June 2024. Available at: https://www.microsoft.com/en-us/research/blog/introducing-aurora-the-first-large-scale-foundation-model-of-the-atmosphere/
[2] Medium, “Chai-1: revolutionising drug discovery with artificial intelligence — a comprehensive review”, September 2024. Available at: https://medium.com/stxbp1-insights/chai-1-revolutionising-drug-discovery-with-artificial-intelligence-a-comprehensive-review-f783c77b9400
[3] Boston Consulting Group, “GenAI Doesn’t Just Increase Productivity. It Expands Capabilities.”, September 2024. Available at: https://www.bcg.com/publications/2024/gen-ai-increases-productivity-and-expands-capabilities?utm_source=substack&utm_medium=email
[4] Ibid
[5] Deepmind, “AlphaProteo generates novel proteins for biology and health research”, September 2024. Available at: https://deepmind.google/discover/blog/alphaproteo-generates-novel-proteins-for-biology-and-health-research/?utm_source=substack&utm_medium=email
[6] NVIDIA, “NVIDIA Expands Large Language Models to Biology”, September 2024. Available at: https://blogs.nvidia.com/blog/bionemo-large-language-models-drug-discovery/
[7] Bain & Co, “Technology Report 2024”, 2024. Available at: https://www.bain.com/insights/topics/technology-report/?mkt_tok=Mzc4LU5ZVS0yMjAAAAGV7NNzRZ7ctIZ5fy7ihZTgwykVHJskPcK_zpoK8vBUr9Zh0ef0xHysaQd42ezFBKzzWX6svQ4NFG40aJzdtlUtNl9VpLoN6in8EaZlZCyUkxFU&utm_source=substack&utm_medium=email
[8] Ibid
[9] SSRN, “The Effects of Generative AI on High Skilled Work: Evidence from Three Field Experiments with Software Developers”, September 2024. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4945566&utm_source=substack&utm_medium=email
[10] Amazon Web Services, “”, August 2023. Available at: https://aws.amazon.com/de/blogs/germany/amazon-codewhisperer-kostenlos-fuer-den-individuellen-gebrauch-ist-nun-allgemein-verfuegbar/
[11] Griproom, “The Dawn of a New Era in Tech: Tracing the Historic Shift from Labor to Computing Power in the Tech Industry”, 2024. Available at: https://www.griproom.com/fun/google-reportedly-now-spending-more-on-computer-power-than-people?utm_source=substack&utm_medium=email#google_vignette
[12] Data Centre Magazine, “How Data Centres Can Make the Global AI Race Sustainable”, September 2024. Available at: https://datacentremagazine.com/technology-and-ai/how-data-centres-could-make-the-global-ai-race-sustainable