Why AI won’t take your job just yet | New Scientist Weekly podcast 234

TL;DR

A new study shows that the adoption of AI for automating human work is currently too expensive for most businesses, leading to a slower rate of job displacement than previously predicted.

Transcript

hello and welcome back to New Scientist weekly your curated selection of the week's science stories I'm Timothy Revel in New York and I'm Christy Taylor also in New York this week on the Pod how fungi and soil grow faster when you play sounds at it how to grow chickpeas on the moon and some new data on how people are being affected in US states wit... Read More

Key Insights

• 👨‍💼 Previous studies focused on the technical capability of AI, while this study examines the cost-effectiveness for businesses.
• 🛩️ Computer vision tasks make up a small part of most jobs, with the proportion of tasks per job that can be automated being relatively small.
• 😚 The impact of AI automation on human workers' salaries and benefits is more significant than the number of jobs lost.
• 👊 There are other types of AI, such as chat GPT, that have already been shown to impact job availability and earnings for human freelance workers.
• 🇨🇷 Cheaper AI deployment and development are being explored, but the financial and environmental costs remain a challenge.
• 💨 New research suggests that playing sounds to fungus can stimulate faster growth, with potential applications in restoring degraded areas.
• 👂 The impact of sound on fungi is not yet fully understood, but it may involve an energy input mechanism similar to the piezoelectric effect.
• 😋 Chickpeas have been successfully grown in simulated moon dust, offering potential for sustainable food production in space.
• 🤖 Challenges remain in terms of latency and fragility, but researchers are optimistic about the future development of robot avatars for remote experiences.
• 😘 The temperature at which bacteria are grown can significantly impact their predatory behavior, with lower temperatures transforming prey into predators.

Questions & Answers

Q: What was the focus of the study on AI automation?

The study examined the cost-effectiveness of using AI technology, specifically computer vision, to automate human work tasks.

Q: How many businesses in the US could cost-effectively automate tasks using AI computer vision?

Only 8% of non-farm businesses in the US have tasks that could be cost-effectively automated using AI computer vision.

Q: What proportion of US worker compensation would be affected by the automation of vision tasks?

Just 0.4% of US worker compensation would be affected by the cost-effective automation of vision tasks.

Q: Is the adoption of AI technology for automation expected to be gradual or sudden?

The study suggests that the initial wave of automation through AI may be more gradual, resulting in less of a shock to human workers.

Summary & Key Takeaways

• A study focused on the cost-effectiveness of AI technology for automating human work, specifically computer vision tasks.

• Only 8% of non-farm businesses in the US have tasks that could be cost-effectively automated using AI computer vision, impacting just 0.4% of US worker compensation.

• The initial wave of automation through AI may be more gradual than anticipated, with less of a shock to human workers.