Driving a train is well on the way towards being a job that people used to do, striking unions is only accelerating the process.

Funeral Directors? Healthcare workers and the most obvious one

Politicians (they wont vote themselves out of a job).

One more point

Being somewhat elderly, I prefer the older version of AI where all that was required was a nail on the back of the stable door for the vet to hang his trousers on...

A stairway to Heaven and a Highway to Hell give some indication of expected traffic numbers

Knowing a reasonable bit about this area

AI is a very real issue, not just because of one scenario, but because there are so many potential scenarios and areas of life where it will have both a positive and negative impact.

ChatGPT has helped make this more of a layman's issue now, but it's not like it's suddenly happened overnight.

AI is used today, in thousands of practical applications. People just aren't aware of it. It's a creeping influence that is creating dependancies that once activated are very hard to wind back, without serious consequences.

For example it's used extensively by credit card companies to help determine if someone is a higher or a lower credit risk, often coming up with models that can provide surprising answers, such as you're less likely to be a risk if you play tennis. This is a real example, btw, and that was five years ago. It's also been used heavily in the accountancy business for years... if you're early in that field or thinking about making it a career working for the bigger firms, I'd suggest re-evaluating that choice.

On a general level, some of the main issues before you even get into the benign vs. threat scenarios are

• Creeping automation leading to slow burn mass unemployment
• Mass unemployment leading to resentment, pushback, clashes etc
• First mover advantage – the race isn't really about improvement for society, it's about absolute power and wealth. Crack AI and you are potentially unchallengeable
• It isn't just Google and Open AI. It's a global competition – China could win. What friction would that then create?
• Lack of effective global oversight or regulation
• Rapidly developing skills gap – people who can use it effectively to augment their jobs vs those that can't (already happening in accountancy)
• Bias, both in development decisions and datasets
• Over-dependence – losing socially valuable/critical skills as we transition to more and more AI automation
• Various economic scenarios where someone decides that it should run the markets leading to unintended negative consequences eg. systems collapse

There are way more but you get the idea. Basically I find it endlessly amusing that these people talk about planning, regulating and controlling something that will end up thousands of times more intelligent and faster learning/improving than them, at which point even the best laid plans will immediately become useless.

As for benign vs threat, assuming we reach AGI and then the Singularity the terminator scenario is a red herring. There are several far more worrying (and higher risk) negative outcomes, such as

Giving the AI the wrong goals, or it deciding to set its own, leading to unintended negative consequences
Not being aware when we reach AGI, allowing it to hide itself and improve, finding out when its too late etc
The runaway / intelligence explosion scenario – that it eventually gets to a point where it's intelligent enough to work out how to self improve, rapidly
That we are simply unable to communicate with it or comprehend it, and it us. The oft cited example would be the human / ant relationship
The fact that even if it is benign, it might be set a goals that sound beneficial but actually end up dystopian, eg. 'world peace' might involve it coming up with controls and ways to prevent conflict which could lead to social restriction or more simply social collapse, depending on what it's interpretation of 'peace' and 'conflict' might be

And so on... A lot of these are way more likely than the reaching out into the physical world / terminator scenario.

I highly recommend reading these two books on the subject. One is more accessible, the other harder to get through but both leave you with a very rounded idea of the benefits/negatives/likely paths and explain the above way better than I ever could.

Easier read – [Link]
Harder read – [Link]

Edited by SW19 CPFC (01 Jun 2023 2.29pm)

Did you know? 98.0000001% of people are morons.

here is an idea.....

the more sophisticated the Financial system....the more volatile.

The derivatives markets have been implicated in previous crashes.

Ancient Rome had low inflation and few surprises....mostly.

AI could create a Frankenstein-Terminator scenario in the financial markets.....leading to the Mother of all Crashes ?

Eze Peasy at Anfield....

A cool post, interesting and a little bit scary....Hal in Kurick's 2001 was scared at the end and now we are.

'Who are you and how did you get in here? I'm a locksmith. And, I'm a locksmith.' (Leslie Nielsen)