Reverse AI psychosis (not promoting AI)

I empathise with this feeling. I suspect companies benefit from analysis paralysis and media coverage, so they are all the time making extreme claims; and the media echoes it.

Below I consider four categories and their relation to AI.

I find it's a useful exercise one can undertake to frame the problem, to partly disentangle these aspects, and also to partly overcome analysis paralysis.

News

Mainstream media reporting on AI is hard to watch or read. Underlying it all, the boomerism and doomerism, there seems to be a theme where they are enjoying the ride. After all, attention is all they need, as their revenues will increase either directly or indirectly.

Conversely, there are real aspects of AI that impact our lives, and the extent to which it will be accepted is unclear. And on this most pragmatic area, there is less reporting.

Cognitive abilities

The cognitive capabilities of the systems remain unclear (they may have none, depending on definitions).

But they can automate certain tasks, and a society as ours will, on average, happily take these, despite of ethical concerns.

There is also little emphasis on their biases, limitations and so forth (in relation to hype). I find these worth reading about.

Environment

I wouldn't be too pessimistic about people's awareness of the harms. 

• AI fanatics will get large bills, so efficient AIs will eventually take over (and, they may consume a more reasonable amount of energy).
• Also as green energy expands, and we adopt it, the systems will end up running in those networks. Though at the moment, that's not how it is.

Poorer countries mine certain rare earths and metals. This invisibility (unless one searches it) leads many to careless purchasing of electronic devices.

Data-centres though, due to energy demands, are built in rich countries (and will affect poorer neighbourhoods the most). Data-centres do have real impacts on daily life, and many are being rejected (all over the modern world).

This is a place where we can raise our voice (e.g. writing to MPs).

Job and Purpose

The lack of freedom to choose AI/not-AI at work is indeed worrying. I reject to be forced to use it, and didn't take a job offer for this reason. Also why if it wasn't in the contract? But there aren't yet regulations in place (as far as I know) so the only alternative would be a different job either within or outside the company.

But as suggested, you can opt-out of most personal use (outside work). I heard people that say they use it to employers, but they don't, which is funny. This may have consequences though, so I don't advice it.

There is also the philosophical aspect, as part of the job or in general, of purpose: what will we do if AI replaces us? I think slowly we will discover new meanings, but it is a complex situation.

-------------

Disclaimer: there may be incorrect assumptions or inferences, but that's my current perspective at least.

If tech companies were taxed properly then maybe we could have UBI, (universal basic income), where most people wouldn't need to work full time, people could be free to have their own busineses and creative projects.

Yes; I think it's sensible to require:

• Taxing them appropriately ensuring the wealth they built on humanities' knowledge is partly distributed back to society;
• Passing laws that punish harmful actions and require thorough testing (as analogy: what biological labs have to do, including thorough testing, guardrails,..)
• Auditing of their technologies for biases, discrimination: What would be of a homosexual if this tech was trained in the 50s, when Alan Turing was "led to" take cyanide? This experiment can be done (cutting off the training data to 1950) and one would guess it wouldn't be nice by today's standards. So how are the systems now currently biased? (There are many interesting studies on the current biases.)
• Stricter planning for data centres, with approval from local communities. (On this regard, I just found this site https://stopdirtydatacentres.eu/)

Some have argued that their datasets (the digital information they train their systems with) should be documented and public or auditable, even in some general way. I also agree with this.

Seems to me these are aspects that we could press our representatives to speed up on, and possibly have more discussion about, rather than going on with the "inevitable heaven or hell" framing.

(I think several countries are developing "AI acts" but personally have not read anything on it so far.)

There may be many other aspects to add to the list?

Yes; I think it's sensible to require:

• Taxing them appropriately ensuring the wealth they built on humanities' knowledge is partly distributed back to society;
• Passing laws that punish harmful actions and require thorough testing (as analogy: what biological labs have to do, including thorough testing, guardrails,..)
• Auditing of their technologies for biases, discrimination: What would be of a homosexual if this tech was trained in the 50s, when Alan Turing was "led to" take cyanide? This experiment can be done (cutting off the training data to 1950) and one would guess it wouldn't be nice by today's standards. So how are the systems now currently biased? (There are many interesting studies on the current biases.)
• Stricter planning for data centres, with approval from local communities. (On this regard, I just found this site https://stopdirtydatacentres.eu/)

Some have argued that their datasets (the digital information they train their systems with) should be documented and public or auditable, even in some general way. I also agree with this.

Seems to me these are aspects that we could press our representatives to speed up on, and possibly have more discussion about, rather than going on with the "inevitable heaven or hell" framing.

(I think several countries are developing "AI acts" but personally have not read anything on it so far.)

There may be many other aspects to add to the list?