This is always an interesting topic to dig up data on. We are always hearing about automation being an impending (we're almost there!) danger that is accelerating. The data doesn't bear this out, however.

Step 1. determine how to measure automation's effect. The best indicator is an economic term called "productivity". We know productivity and automation are correlated because the biggest shifts in productivity come from technological advance (and not say, superhuman capabilities being developed by worker)

Step 2. Look at historical data. It turns out productivity has been climbing but not as fast as it was from 1990-2010 (very steep climb). In fact the last decade 2010-2020 has seen some of the slowest increases in productivity since the middle of the 20th century.

Source: inferred from this productivity vs compensation graph from world economic forum.

https://assets.weforum.org/editor/HFNnYrqruqvI_-Skg2C7ZYjdcXp-6EsuSBkSyHpSbm0.png

Speaking of corporate lobbying... this guy is from the American Enterprise Institute. I haven't looked them up recently but as far as I remember they are a neoliberal think tank that are funded largely by the massively wealthy with the aim of influencing government policy to reduce social spending and increase corporate subsidies and privatizations.

I can imagine politics and corporate money will be at the core of classroom cost-cutting and student to teacher ration increases.

1. On the one hand, there will be the AI companies lobbying the politicians w/ money and sending sales delegates to the ministry of education to brief them on how to cut costs with their new AI tools.

2. totally unrelated companies, with no connection to education or AI will lobby politicians to cut education funding too ("the students can learn with AI!") because the less the government spends on education, the more that the megacorps can ask for the money to be used on corporate subsidies and corporate tax cuts.

Building on my aforementioned comparison of ChatGPT to a moderately skilled tutor, there is a downside to AI tutors too: Once the ministry of education figures out that most students have robo-tutors, they might be tempted to use it as an excuse to increase class sizes.

How it will affect "education" specifically: I think if we narrow the question down to AI LLMs like ChatGPT specifically, then the effect will be that of having access to a moderately skilled tutor. Ask the right questions and it can teach you how to do a task quite well!

A wonderful truism. It is always important to get things right. But I'm not sure if the author has actually discussed substantively about how exactly we might go about doing this right.

This is certainly beyond the point of the article, but I'd like to chime in that the heuristics we use to determine allocating funding based on need are often wrong. When I was in university, some wealthy students were able to receive educational grants simply because they technically met the criteria on paper, whereas the spirit of the grants were clearly to reduce the burden on economically disadvantaged folks. A sworn declaration and an interview to confirm material need should suffice. I can see AI further entrenching the use of unrelated qualities as a (potentially incorrect) heuristic for deciding whom to assist.

I'd say two-way communication, and planning are some of the shining strengths of the tech.

As for creativity and problem-solving ... it depends. The kind required for many white-collar jobs is covered by ChatGPT. We generally have far less leeway to act independently at work than we think.

The higher level C-suite stuff though? Maybe it's not replacing their jobs. But then again, isn't it their assistants that do their work for them anyway?

Sadly for software engineers, this is the latest competition for their job stability after the ongoing huge influx of university graduates in the STEM fields, and very skilled people trained abroad. Junior engineers and recently immigrated engineers work for cheap. AI works for free.

When ChatGPT (or something similar to it) automates construction, hospitality and tourism (e.g. culinary, hotel workers, etc), manufacturing, transportation, call me up so you can call me out.

I compltely agree with architecture and distribution & logistics (the planning part of it anyway).

Author: ChatGPT writes flawless computer code.

Also author: Information Technology has a low risk of automation.

Me: raises an eyebrow.

Sadly the current Ontario system is a major culprit here. Kids in grade 10 have to make a choice between the O/M/C/U streams that will affect them for life. Don't have a good chemistry tutor at age 15? Great, you're never going to uni unless you make some intense sacrifices down the road to re-do your high school credits, and be willing to give up any college credits you've accumulated in the transfer.

I can't make sweeping conclusions about a disruptive technology like this. We don't know the consequences, let alone the specifics of what strategies to use to shield ourselves from the sequelae.

Hard disagree with the conclusion. Computers in general, and ChatGPT in particular, has been historically VERY good at replacing white collar and bureaucratic labour. It's particularly terrible at replacing manual labour, skilled trades and human-to-human emotional interaction. The axiom that low-education jobs are getting automated is an axiom we've been asked to accept for decades but it's just not happening.

My wife "befriended" it and just had conversations with it as if it was a new friend and conversation partner. It was fun seeing her "socialize" with it for a few weeks when it came out. Everything from the meaning of life, consciousness, religion, philosophy, instructional guides and then some were discussed.