I feel I should say something about artificial intelligence.
To some degree I am hindered when contemplating doing so by the fact that I am bemused by all the fuss about it.
I was an early adopter of IT as a business tool. I bought my first serious personal computer in 1984. The accounting practice I ran from 1985 never knew a non-IT based existence.
Similarly, I embraced networking, email, the web, and more as each became available. Some (most especially networking and email) had significant impact on working practices.
In the late 90s I was involved in two dot.com companies, one of which was quite successful, and both of which made a return for their shareholders.
As blogging, Twitter and much else has happened I have also embraced those changes. Again, each has radically altered work practices, not least mine.
And now the functionality of many of the tools used to date has been enhanced by, in essence, automating the intermediate instructions between defining some tasks and completing them. Not that there is anything desperately unusual about that: real world robotics seems to have been moving in this direction for some time. So too has gaming.
So what is the fuss about AI about when so much change has already happened in my lifetime and I have been discussing AI since the mid 90s, at least, when I had clients engaged in the field?
To some extent, I again stress that I am not sure what the concern is, most especially when I remember not dissimilar paranoia about the web in the 90s.
To date IT has not created mass unemployment, even though it has fundamentally changed many jobs.
Likewise, the extraordinary power of the web has not rendered human thinking redundant.
Nor has IT, as yet, fundamentally changed human relationships, although the onset of decent high-quality headphones at affordable prices did mean I did not spend my sons' teenage years yelling at them to turn the volume down.
So, I am not panicking. But, that said, I do realise that AI creates risks because it replicates human skills, as IT has always done.
I also recognise that this means that some jobs are at threat. I also know vast amounts of work is currently not done: there are no shortages of opportunity for gainful work in society.
And, of course I recognise the risk from ‘super-intelligence', most especially within politics, where ‘normal-stupidity' is commonplace.
More particularly, the risk of further concentration of economic power in the hands of a few corporations is especially worrying.
The continued absence of effective means to properly tax IT companies might become an ever-bigger issue.
That will be exacerbated by the growing need of government for revenue, most especially as it becomes the major source of new employment in the essential public services that will be the real foundation of this new economy.
I also see the risk to the climate change agenda if AI absorbs vastly excessive energy, as Bitcoin does, without any net gain to humankind.
So, when it comes down to it, AI is all about a classic political economy power struggle. Corporations will seek power. Government must constrain them. Tax must be collected, probably in increasing amounts. The dependence of people on the state as the need of the private sector for the skills they have declines. This will create tensions.
And in all that the biggest challenge is to our ability to innovate ideas, and not to our intelligence (the two not being the same thing, as some universities seem determined to prove).
Can we rethink the relationship between the state and private sectors?
Can we imagine a courageous state, rising to the challenge of creating employment to meet need?
Might tax be appropriately reformed?
Can the climate transition happen despite AI, which appears to be inherently energy intensive?
Might we learn how to imagine and counter AI threats?
Can crime detection morph in an era when fraud will become much easier?
Can we adjust our ideas of liberty within these constraints, recognising that machines (and their owners) do not enjoy a special claim in that regard?
None of these are, I suggest, questions AI is able to answer - because they require imagination that I do not think it capable of. In essence, they go to to the soul of the issue, which is a domain to which I think only humans will remain privy.
And I may of course be wrong. But, we have adapted so many times. I think we can again. I live in hope, whilst recognising the challenges.
Thanks for reading this post.
You can share this post on social media of your choice by clicking these icons:
You can subscribe to this blog's daily email here.
And if you would like to support this blog you can, here:
Apparently Elon Musk thinks AI tools will eventually take over all jobs so humans need only work if they want to. Which made me wonder who or what would be looking after him in a care home. And indeed how the rest of us who are not tech billionaires will afford all of these wonderful AI services.
The sort of change that we saw with the mechanisation of agriculture and then advanced manufacture (far fewer repetitive and physically demanding rural and urban jobs as human labour was replaced by machines) and to administrative tasks (computerisation replacing filing clerks and the typing pool etc) is coming to more intellectually demanding and creative jobs.
We will no doubt adapt – new jobs will open up as the old close, car mechanics replacing ostlers for example; or website designers, social media influencers, etc. And telephone sanitisers perhaps – but the process may be painful for many.
Adaptation is always painful
But in the mdoern lifteime it has become inevitable
Elon Musk needn’t worry about life in a care home. Rishi Sunak will always be there as a willing supplicant.
Talking about thinking:-
“It is difficult to believe that an economy with below-4% unemployment can sustain this pace of job growth,” Dean Baker, senior economist and co-founder of the Center for Economic and Policy Research, wrote in commentary this week.
https://www.theguardian.com/business/2023/nov/03/us-added-150000-jobs-october-growth-slows
Conveniently forgetting the Second World War US economy:-
https://shec.ashp.cuny.edu/items/show/1510
Almost all of what is currently called AI doesn’t think at all. It finds patterns in data and is able to create new writing/images/video/music that has similar patterns. Since a lot of our “thinking” relies also on first seeing patterns, it looks a bit like what we do some of the time. But actual human thinking involves deduction, induction and intuition, which we can do based on not much experience, and none of which current “AI” does. Think of it like a really flexible database – potentially very useful, potentially dangerous if not regulated properly, but hardly any kind of existential challenge (unlike our inability to socially organise ourselves for our own good, which is leading to climate- and energy-mediated chaotic social collapse).
Agreed
I have major concerns about AI even if some workable and enforceable form of regulation can be agreed and implemented. The digital age has already facilitated huge increases in fraud and AI has the capacity to make fraud harder to identify (voice identification, social media personal data etc). Fraudsters operate outside the law and detecting their activities will be harder once they start using AI for their scams.
The digital age also undermined the financial structures of the creative arts before further damage was inflicted by Covid and lockdowns. For instance a basic fact of life for most musicians was that studio quality recordings were essential to get bookings. Studio time, manufacturing costs and copyright fees (where applicable) were expensive but this cost was worth paying to get the work. The upside was that CDs could be sold at the gigs in the hope that, over time, enough income would be earned to cover the original outlay. However the ease of accurately copying CDs and downloading streamed recordings very swiftly resulted in catastrophically reduced CD sales. The royalty fees deriving from online plays are tiny: Youtube pays artists $0.00164 per stream for their own channel’s streams and $0.002 per stream on YouTube Music. Artists earn a minimum of $0.00069 per stream for music played on other creator’s videos, under Content ID instances. Spotify pays only marginally more. Then Covid came along and in its wake huge numbers of venues never reopened. This affected grass roots clubs and niche genres like jazz disproportionately.
AI brings the threat that copyright enforcement, already weakened by sampling and the ease of copying digital files, may become impossible. Just look at the new Beatles recording just issued. A cassette demo recording with dodgy sound quality of John Lennon singing an unpublished song has been doctored by AI to sound like a recent live studio recording with full production values and band accompaniment. That is all above board as it’s effectively an “in-house” Beatles-managed project, but the music business has always been notorious for copyright infringement and straightforward theft of other people’s work. How will copyrights be protected if AI is used to extract and/or alter individual voices or instruments for use in other settings? Who is going to police that kind of behaviour? It will be harder than ever to detect and history tells us it’ll be tried.
Thanks
Excellent points
And what it sys is live music – real human experience – matters more than ever as tech destroys the apparent value in much else
Richard wrote “….what it s(a)ys is live music – real human experience – matters more than ever as tech destroys the apparent value in much else.”
I agree 100%. I had a career as a musician lasting over 60 years and witnessed the gradual commodification and monetisation of music. The outcomes of that process include the replacement of musicians by disc jockeys, the marginalisation of entire musical genres simply because they didn’t generate enough revenue for record companies, radio & TV stations (and all the ancillary activities that depend on music for their livelihood).
The decision of the Musicians’ Union to accept DJs as members led to the widespread resignation of members (myself included) who had lost much of their income to the DJs and, as night follows day, theatres increasingly use pre-recorded music rather than actual musicians. Friends whose livings largely depended on West End shows in London now find themselves teaching to make ends meet. As the opportunities to make a living out of playing music diminish, it also impacts young musicians graduating from conservatories and colleges: most of them come into the profession to play and now find themselves teaching instead.
Over my 60+ years as a musician, with (fortunately for my family) a parallel career in accountancy which took me to 4 continents, I came to the conclusion that the UK does not value Culture: the overwhelming majority of creative artists scuffle to make a living and need employment outside the arts just to meet basic needs. The austerity-driven decision of the Tories to slash culture funding by 50% in England is drastic in itself, but it also affects the devolved nations since the Barnett Formula calculations for their Block Grant funding is based on government expenditure in England. When a nation doesn’t protect its culture it loses its soul and my conclusion is that’s happening in the UK right now as a result of political indifference and blind adherence to false ideologies.
Another legacy of my musical career is my understanding of what music is. I see it as ‘language’, with all its multiple genres worldwide as ‘languages’ and ‘dialects’ in their own right. Language communicates verbally, but music communicates emotionally and the abstract nature of that means that different people will have different interpretations of the meaning and mood of any piece (which is part of music’s allure). When music doesn’t connect emotionally with its listeners it loses its purpose, hence the importance of live performance. That’s not to say recorded music can’t connect emotionally (just listen to Maria Callas, Louis Armstrong, Jean Redpath, Bessie Smith and millions of others on record – the emotional intensity is obvious), but live performances are unique and of the moment; often they’ll never be replicated and therein lies their magic.
I agree – and know I do not see enough live music these days.
Nor do I play much now. My clarinet lies largely unused, as do five recorders which I used to much enjoy, and still do sometimes, however underrated an instrument it is.
There is something really important about the recorder. I know you can be musical without reading music, but being able to read it does really help, I think. Translating notes from the stave to the instrument can be profoundly mechanical, but when you begin to understand why music is written the way it is, and what the composer wanted to say – which reading a score can help achieve – then I think it helps. It’s a basic form of literacy. I am at no more than a basic level, but appreciating the difference between major and minor matters. Understanding that even then some keys speak in particular ways – with D flat reputedly the saddest of them all – seems important. It puts you on a wavelength with the music, whilst letting you also appreciate it more.
My mum was a primary school music teacher after she was a nurse, and I wish there was more emphasis on that now.
Many years ago,I read an extremely short SF piece about computer intelligence.
This said that all the nations in the world decided that they would link all of their computers together into one giant computing system.
They then achieved this, and posed it the question ‘Is there a God’?
A booming voice said ‘THERE IS NOW’.
oldrafman
Very good
“Can the climate transition happen despite AI, which appears to be inherently energy intensive?”
Timely de-carb & mitigation is unlikely to occur without A.I. input. For the most part A.I. is implemented in data-centres of various sorts. These can be described as both data-processing centres & heat engines (about 95% of the elec that goes in comes out at +/- 36C hot-air – although this is likely to rise to 60C with new types of processors).
Ireland is a popular location for data centres (powered by wind etc). Ireland is also a major importer of water from Spain (as fresh veg and fruit).
Most data centre operators are happy to give heat away for nothing (I know cos I have spoken to them) . Thus for Ireland and other locations the issue is how to implement heat re-use – green houses being perhaps an obvious application given many Irish data centres are located in the countryside (e.g. Apple’s).
Designing/implementing an energy re-use project is mostly rules-based. A.I. works best in environments with well defined rules. Thus one can start to see paths forward for both de-carb (via energy re-use) – and resilience (grown your own rather than import from water poor locations) . Some experiments have started – but it is all very cottage industry stuff with everybody in permanent re-invent the wheel mode. A.I. could & should make a difference.
One area where A.I. could change employment: notary publics – don’t need them anymore with a fit for purpose A.I.
Very good
“One area where A.I. could change employment: notary publics – don’t need them anymore with a fit for purpose A.I.”.
Well, you have just blown the single USP the City of London possesses.
🙂
(I’m so sad).
I have followed the area for many years having started working life in the IT industry, whilst claiming no deep expertise. ‘Computer intelligence’ has been around in different forms going back years. More recently I have come to know the chair of the APPG on AI who kindly shared their reading list – I can share it with anyone who is interested. Excellent stuff from the likes of Hannah Fry, Cathy O’Neill, Anthony Seddon, the Susskind’s, Rana Faroohar which I’d recommend for non-technical audiences.
As has been said about so much technology claimed as the latest silver bullet, we tend to overestimate its short term impact and underestimate its longer term impact. That was true for instance of on-line shopping where the long term consequence is the massive monopoly that is Amazon with its impact on small retailers and suppliers – monopoly and monopsony. Then of social media which was widely welcomed until we saw the dark side in the elections of Trump (and others) and Brexit.
It’s a big topic but I’d flag two big concerns; firstly the very real threats to democracy and political and social stability more widely as AI takes the abuse of social media to a whole new level. It is already happening with evidence of it appearing in the Palestine-Israel conflict. Secondly the deeply monopolistic nature of the firms dominating the sector, especially the combination of tech and finance. We saw this writ large with the interview just held between Sunk and Musk, two people who epitomise the amorality if not outright immorality of their sectors. Utterly lacking in any concern for or appreciation for the ethical, moral or social dimensions. The worst think we can do is to leave it to the techies who with all due respect tend to lack those characteristics, along with the finance folk they align with. See Musk and Zuckerberg for starters.
Biden’s government does at least show signs of appreciating the problems and are leading the way, with the EU not too far behind. Whilst the UK has a strong presence in AI development, with Sunk and co and charge it will be a bit player in regulation and given Sunak’s character that is probably a good thing.
None of that is to dismiss the very real potential of AI in well defined areas such as cancer scans, with well defined data sets and algorithms that can be tested and monitored.
Thanks
I accept your threat analysis
The 2021 Reith lectures were on the topic ‘Living With Artificial Intelligence’ and were given by
Stuart Russell, Professor of Computer Science and founder of the Center for Human-Compatible Artificial Intelligence at the University of California, who seems to know what he was talking about. There are still available https://www.bbc.co.uk/programmes/articles/1N0w5NcK27Tt041LPVLZ51k/reith-lectures-2021-living-with-artificial-intelligence
By far the most alarming was number 2 AI in warfare.
I’m informed by the Work of John Seddon and his crew at Vanguard Consulting, whom I have met, listened to him talk many times and read all of his books. Seddon is an interesting man who I have a lot of time for.
I think the problem with AI is that I don’t think it can cope with requisite variety as well human beings can. A lot of Seddon’s work looks at customer service modules in public and private sectors that are automated and picks up that there is a lot of ‘failure demand’ – people who just don’t fit the algorithms narrowly defined parameters, give up and want to speak to a real person.
To me, Seddon’s work is a timely reminder that only human beings and our special quality of empathy can solve the problems of other human beings. There is a use for AI I’m sure, but it needs to be relegated to lower tasks in the service sector as far as I am concerned.
Those pushing AI, who want lower wage bills, lower taxes, higher profits etc., are far too vocal in my opinion.
I also believe that most AI and indeed IT is just someone’s version of what something should be like and not necessarily the best version either. Yet they’ll tell you it does this, does that blah, blah blah. And my biggest fear is inbuilt bias.
We should be very careful going forward, we should be more sceptical about AI, and its use should be regulated.
Some more thoughts on AI…………..
Today I travelled ‘oop North’ to see my ailing father in law – all of 85 years old and with Alzheimer’s. This week, he lost his mobility and is about two weeks from going into a home after waiting far too long for a diagnosis. He is now bed-ridden and dozes off but like the sun occasionally bursting through a grey cloud-ridden English winter day, he has all too brief periods of cognition.
So, it’s been a tough day, but also a day that validated being human with some wonderful moments.
Firstly we were just there for his partner who has done a wonderful job of looking after him but now is at the end of her tether. She needs to talk about her situation, how she feels, the contradictions, where she feels she could have done better, the guilt of having to find him full time care, the successes , what they should have done and not done, dealing with Adult Social Care, reassurance – so we just listen and take her through it, let her let it all out.
The Friday before, the doctor had called around and it turned out that he too had a father in the same situation, and empathy and information was exchanged and gratefully received about the best performing homes, where to go, what to look out for etc.
When we arrived the community nurse was there helping set up a piece of apparatus that would enable my father in law to make his way to the loo. Yet more time was found to listen, share and pass on good practice and other tips (using medicine cups to drink liquids from for example).
Then a grocery delivery came and a discussion started with the deliveryman talking about his father who had dementia with my father in law’s partner. So, dementia and Alzheimers became social levellers for a minute or two as it brought together a working class Geordie with middle class ex head of Education and Children’s services in the kitchen with a common human experience. The notion that they were not the only ones going through this was smashed, some sort of connection had been made – that someone else really knew what it was like.
And then there was my father in law himself, who with kind words and coaxing (especially when the word ‘pudding’ was mentioned bless him) was able to briefly smile at us and hold our hands intermittently like old friends would.
So, I reflecting on this when we returned home I wondered if AI could have generated the same outcomes that were achieved today?
I think not.
Human beings – turned on, tuned in, taught to care, empathise, love, listen, seeing common cause with others (and decently funded) are pretty bloody unique and awesome. We are nowhere near obsolete folks let me tell you. No way!
I just hope AI is put to work where it is right to put it to work and no more. It can complement us perhaps but it will never replace us. And if it does, we risk accepting second best if not worse.
Chew on that Elon Musk. And since ‘Twitter’ has is nowhere near the value it was before you got your mitts on it, why would you be an expert on anything, let alone AI? Honestly!
Thanks
Good luck to your father in law and all involved
I’ve been there with my late dad
Thank you PSR. A reminder of what humanity is about. And what will not be replicated anytime soon.
Musk and Sunak in their interview showed how as personalities they are little more than bots, devoid of humanity. The very last people to understand let alone regulate AI.
Declaring my interest. I am an aging software engineer with an interest in AI
Will AI take our jobs? Will a superintelligence become our overlords? Maybe. I think that is far in the future.
I think the real clear and present danger here today is the ability to use AI to create Deep Fakes of podcasts and youtube videos. Showing public figures saying and doing things that are not fact. You cannot trust anything you see or hear on the internet. Did they say it or was it manufactured? Prove it!
The ability of the nefarious actor to manipulate public opinion and affect elections is here now and is incredibly dangerous
I agree – this will be worrying
How is AI going to engage with out fundamental problem that we are a prosocial species but allow our lives to be stunted even destroyed by ways of doing things that aren’t prosocial or fully prosocial.
An example of the former is the failure to support a cease fire in the Israeli/Palestinian conflict with the view that a permanent settlement and rights agreement is required.
In the latter we have the pretense that free market capitalism is by definition always acting in the interests of the many.
Can machines think? I doubt it. They are just really good (read fast and getting faster) at computational statistics. The algorithms which inform their calculations and allow them to construct their simulation of reality, which we mistake for intelligence, are riddled with human bias. They reflect some of our worst prejudices. The clear and present danger lies in allowing so-called AI to make decisions on our behalf both individually and collectively as a society and to assume that those decisions are both rational and fair. My big worry is that this is increasingly happening – computer says no!
Another aged software engineer offers a comment…
We have now reached the point at which decisions reached by AI cannot be either validated nor understood by a human being. If such systems are allowed to be widely applied to anything that affects the welfare of human beings (the law, civil rights, and worse, war) we are trully on the road to destruction. AI is not another aspect of the industrial revolution – it is a totally different order of magnitude of change. That is why legislation is imperative.
We should be very careful about letting AI do our thinking for us. It may be faster and more accurate. It may be able to relieve us of mundane tasks. But who decides what is mundane? Why is it always better to go faster and eliminate errors? Mariana Mazzucato has written about the problems of outsourcing work from public services to contractors – the skills drain, the eventual dependence on the service provider – we risk the same outcomes if we outsource our problems to artificial intelligence.
Well put
It is possible to make nice intellectual distinctions that demonstrate all kinds of matters can only be understood, or at least performed by human beings. True and important, but I think it overlooks an important point. That is, what is really important to human beings that could be handled without any input from humans, save as wholly dependent consumers; from necessity?
Money. Money is binary, and easily digitised. Money is in the AI wheelhouse; front and centre. Money enables, manages, controls everything. That, I submit is what this is really about, and why the lights of Big Tech and Digital Capital are all lit up. I have always said; that will turn their mind to banking as the next stage in their development of global hegemony. They know the routine. Begin with convenience. Humans fall for convenience, instant gratification; immediate satisfaction of every idle want. We are the turkeys being fattened for Christmas; someone else’s Christmas. We already know we will wake up only when it is too late. We always do; because we never, ever look back and see the carnage that has been wrought just bringing us to where we are now. We think it all fine, because we never look back, perhaps because we know enough to fear what we would see; about ourselves.
Cynically appropriate, John