(More on Mega-trends - 2022 - cont. 6)

6. Increasingly Autonomous (artificial intelligence - AI)

Introduction

"...astonishing improvements in the ability of software and machines to solve problems and perform complex tasks without explicit human guidance......How AI and related science research and technology capabilities are helping to boost productivity and solve humanity's greatest challenges and socio-economic considerations..."
CSIRO, 2022

Can machines think?

Impacts of this megatrend

i) increase research into artificial intelligence (AI)

The size of the global market, including R&D, in AI is increasing, especially in China.
"...The global growth in AI's research and investment is likely to continue and transform how all industries operate in the coming decade..."
CSIRO, 2022

 ii) new possibilities provided by AI advances

Already AI's performance exceeds human benchmarks in areas like image recognition, natural language processing, etc.
"...recent advances in generative models means that AI systems can now generate synthetic text, audio and images to a level that is indistinguishable from non-synthetic outputs.

These techniques are increasingly being applied to generate synthetic datasets, which can be used to train new algorithms without the need for real data. Generative approaches can provide researchers and developers with access to large, representative datasets that can be difficult or unethical to collect..."
CSIRO, 2022

However, these technologies can be used for malicious purposes such as creation of fake impersonation videos (deep fake).

iii) enabling new discoveries

"...AI is transforming the speed, quality and breadth of practically every field of physics, nature and social science and creating new paradigms of knowledge discovery..."
CSIRO, 2022

Some examples include predicting 3-D structure of proteins so that they can identify the protein's function and speed up the development of new treatment, conduct scientific experiments and discoveries that are indistinguishable from the best human scientists.

 iv) improving computational power and quantum computing

Purpose-built computer processors are being developed to handle matrix algebra and improve the efficiency of machine-learning; quantum computing will substantially increase computational speed and efficiency, especially for AI and machine-learning systems.

v) a growing portfolio of useful artificial intelligence applications

Some examples include

- use a machine learning to simulate wildfire movement and support effective and efficient fire suppression activities (CSIRO's Spark)

- advanced flight-planning which applies machine learning to achieve faster, smoother and more fuel-efficient flights (Constellation)

- using computer vision to detect and alert fatigue truck drivers (Seeing Machine)

- development of airborne drones that can spot sharks in the surf better than a human observer

vi) global investments in research and development

Global research on AI spending is growing faster than GDP. However a majority of this growth is driven by a minority of countries, eg Israel, Japan, Germany and South Korea.

High levels of investment in R&D will help fuel future scientific discoveries, innovation and applications.

There is a lag between how quickly AI breakthroughs can be translated to business and policy applications.

vii) the shifting centre of research and  development investment

Strong economic growth in Asia is driving investments in R&D, with China and Korea dominating. This is expected to continue and will include India and ASEAN.

viii) ethics of artificial intelligence

There is concern about AI's risk to safety, livelihoods and the rights of people. As a result, codes of conduct are being developed.