Neither, seems to be the answer to the question posed by the title. But if one thing AIs do is filter the world for us, the question of who does the filtering and in whose interest they do it becomes very important. As with other free services, free algorithms will be provided in expectation of a benefit to somebody, and that somebody may very well not be the end user. So far so unexceptional (and putting it under the heading of AI doesn’t change the substance of an issue which has been around a good while). But if this is a problem, what are the pressures and processes which will work to relieve it rather than reinforce it? Here, the argument rather fades away: we are told we need clear laws and well-accepted procedures to regulate AI, but there is little suggestion here about what they would say or how we would get to them. It’s slightly unfair to single this piece out for what is quite a common problem: when challenges are technology driven, but solutions need to be socially driven, it’s a lot easier to talk about the first than the second.
This opinion piece gives a punchy account of reasons to oppose the introduction of a universal basic income – which makes it useful as a clear statement of one side of the argument, though at the expense of a more balanced assessment. Interestingly, three of the five reasons given are to do with work incentives and the role of work in supporting self-worth. In the context of the wage-based economy which has dominated industrial societies that may make sense – but prompts the interesting further questions about whether they are universally applicable and will survive the next rounds of work automation.
This post is just a teaser, but a teaser for something interesting in both style and substance. Starting next Monday, we are promised a five day intensive course on artificial intelligence. AI is for many people at the stage where there is lots of fragmented insight and understanding, but little which brings the fragments together to form a coherent whole. So there is a gap here well worth filling – and this looks to be a neat low key way of filling it. Watch that space.
If you had to write down a list of innovation methods and techniques, how many could you come up with? However long your list, it’s a fair bit that it won’t have as much on it as this landscape of innovation approaches (also available as a more legible PDF to cut out and keep).
Methods are grouped into four overlapping ‘spaces’. There’s room for debate about what best fits where and there is a broad range from mainstream to eclectic – but that in itself is a good start in challenging assumptions about methods which appear natural and obvious and indeed about the kind of innovation being sought.
A recurrent criticism of governments’ approach to digital services has been that they have been over focused on the final stage of online interaction, leaving the fundamental organisation and operation of government services unchanged. More recently, design has more often gone deeper, looking at all elements of the service and the systems which support it, but still largely leaving the underlying concept of the service in question unchallenged and unchanged. This post takes that a stage further to look at options for the underlying operating model. Eight are set out in the post, but it is probably still true that most government service design and delivery happens under the first two heading. What are the prospects for the other six – and for all the others which haven’t made it onto this list?
This is a follow up to a post covered here a few days ago which looked critically at outsourcing, starting from the fundamental question first posed by Coase on what organisations should do and what they should buy. This second post is at one level a short summary of the first one, but it’s also rather more than that. It puts forward a slightly different way of framing the question, making the point that time and uncertainty are relevant to the decision, as well as pure transaction cost narrowly defined.
There are transactions, which are in the moment, and imply no further commitment or relationship. There are contracts, which are a commitment to future transactions, and depend on shared assumptions about the future conditions in which those transactions will happen. And there are organisations, which exist in the space beyond contractual precision and certainty.
To complete the hat trick, there is also a separate post applying this thinking to Capita. Even for those less interested in the company, it’s worth reading to the end to get to the punch in the punchline:
In important ways, this is the service that Capita provided and still provides: the ability to blame problems on computers and computer people, while ignoring the physical reality of policy
How many design innovation toolkits are there? The answer seems to be that there are more than you might think possible. Over a hundred are brought together on this page, which makes it an extraordinarily rich collection. There are lots of interesting-looking things here, some well known, others more obscure – though it’s hard not to come away with the thought that the world’s need for innovation toolkits has now over abundantly been met.
Being a subversive is hard work. That’s partly because being the odd one out takes more energy than going with the flow, but it’s also because subversion decays: yesterday’s radicalism is today’s fashion and tomorrow’s received wisdom, to be challenged by the next round of subversion. If that sounds a bit like the innovator’s dilemma, that’s perhaps because it is, with some of the same consequences: you can ride the S-curve to the top, but if you don’t flip to the next curve, your subversion-fu will be lost.
The reciprocal effect – which is more the focus of this post – is the effect on the organisation being subverted. Just yesterday, I heard ‘minimum viable product’ being used to mean ‘best quick fix we can manage in the time’. The good intention was still there, as was an echo of the original meaning, but the hard edge of the concept had been lost, partly I suspect, because it had become dissociated from the conceptual context which gave the original meaning. That’s not deliberate degradation but – as the post notes – is the consequence of a virtuous attempt to bring in new thinking, only for it to get absorbed by the wider culture.
So the challenge for subversives remains: how to keep subverting themselves, how to stay one curve ahead.
The idea of the black box pervades a lot of thinking and writing about AI. Mysterious algorithms do inscrutable things which impinge on people’s lives in inexplicable ways. That is alarming in its own right, but doubly so because this is new and uncharted territory. Except that, as this post painstakingly points out, it’s not actually new at all. People have been writing software about which they could not predict the outputs from the inputs since pretty much since they have been writing software at all – in a sense, that’s precisely the point of it. And if you want to look at it that way, the ultimate black box is the human brain, where the evidence that we don’t understand the reasons for our own decisions, never mind anybody else’s, is pretty overwhelming.
The need for precision at one level – software doesn’t cope well with typos and syntax errors – doesn’t translate into precision at a higher level, of understanding what that precisely written software will actually do. That thought came from Marvin Minsky in 1967, but people had been writing about black boxes for years before that, when the complexity of software was a tiny fraction of what is normal now.
The fact that this is neither new nor newly recognised doesn’t in itself change the nature of the challenge. What it does perhaps suggest, though, is that strategies developed for coping with these uncertainties in the past may well still be relevant for the future.
The bigger the underlying change, the bigger the second (and higher) order effects. Those effects often get overlooked in looking at the impact of change (and in trying to understand why expected impacts haven’t happened). Benedict Evans has always been good at spotting and exploring the more distant consequences of technology-driven change, for example in his recent piece on ten-year futures. ‘Cascading collapse’ is a good way of putting it: if the long-heralded but slow to materialise collapse of physical retail is beginning to appear, what consequences flow from that?
Today HMRC announced that 92.5% of this year’s tax returns were submitted online. That too has been a slow but inexorable growth, taking twenty years to go from expensive sideshow to near complete dominance. There is more to do to reflect on the cascading collapses that that and other changes will wreak not just on government, but through government to society and the economy more widely.
Organisations, including governments, follow fashions. Some of those fashions change on short cycles, others move more slowly, sometimes creating the illusion of permanence. The fashion for outsourcing, for buying rather than making, has been in place in government for many years, but there are some interesting signs that change may be coming. One immediate cause and signal of that change is the collapse of Carillion, but that happened at point when the debate was already beginning to change.
This post goes back to the roots of the make or buy choice in the work of Ronald Coase on the nature of the firm. The principle is simple enough, that it makes sense to buy things when the overhead of creating and managing contracts is low and to make them when the overheads are high. The mistake, it is argued here, is that organisations, particularly governments, have systematically misunderstood the cost and complexity of contract management, resulting in the creation of large businesses and networks of businesses whose primary competence is the creation and management of contracts.
One consequence of that is that it becomes difficult or impossible to understand the true level of costs within a contractual system (because prices quickly stop carrying that information) or to understand how the system works (because tacit knowledge is not costed or paid for).
All very thought provoking, and apparently the first in a series of posts. It will be worth looking out for the others.
Issues of data aggregation and de-anonymisation are hardly new, but there’s nothing like a good example to make an issue more visible – and secret US bases revealed through aggregated data from fitness trackers are about as good as it gets.
The real issue though is less such revelations and more the implications for data and privacy more generally. This article argues powerfully that to see this as an issue of individuals and clickthrough privacy policies is to miss a very important point. People can’t consent to the ways their personal data will be used and the risks that carries, because service providers don’t and can’t understand those things themselves, and so can’t explain them in a way which makes consent meaningful. That has some important data policy implications, including much stronger liability for data breaches, and keeping the amount of data captured and held to a minimum in the first place. Those are not new suggestions, of course, so as ever the real question is not how the risks could be managed better, but how incentives can be aligned to ensure that the risks are in fact managed. And that is a political and social problem, not a technical one.
An understanding of quantum field theory apparently demonstrates that in large convoluted organisations, hierarchical structures with one person in charge can’t work, because the level of complexity becomes impossible to manage. That’s essentially the long standing perspective of systems thinking – if you want to change a system, you have to change the system – and while it’s entertaining to see the point made from a different standpoint, the real question is not whether this approach can provide a diagnosis, but whether it can offer a prescription for change.
It’s almost certainly unfair to make a judgement about that on the basis of the transcript of a short radio interview, which is what this is, but what’s striking is how quickly the prescription becomes a platitude. If political decision making were more distributed, as decisions in the brain are distributed between neurons, better decisions would result. That may well be true, but doesn’t get us very far. Part of the suggestion here seems to be a form of subsidiarity, which is a good start, but one big reason politics is hard is because decisions really are interdependent. What we have to do, apparently, is create mechanisms whereby participation translates into actual decision making. Well yes (or at least, well maybe), but asserting a solution falls a very long way short of describing it.
It’s included here despite all that for two reasons. The first is as a reminder that politics is hard and that insights from other disciplines are unlikely to provide magic answers to long standing and intractable problems. The second is that problems of political decision making are long standing and intractable and answers, ideally less magical, are still very much needed.
This is a fairly straightforward tour of the basic income landscape. Perhaps it is most useful for something unintended, drawing out the extent to which the debate on the virtues or otherwise of a basic income is conducted at cross purposes. People use the phrase ‘basic income’ to mean two very different things (probably many more, but two will do to start), which we might call ‘adequate’ and ‘supplementary’.
An adequate basic income is one which is enough to live on, not luxuriously but, well, basically. A supplementary basic income is not in itself enough to live on, but is enough to make a real difference to people’s lives and choices, particularly at the lower end of the income scale. This article concludes that a UBI doesn’t distort work incentives, but concludes that from looking at examples of supplementary basic income. Even if that conclusion is robust, it can’t in itself tell us anything about adequate basic incomes. This piece does better than many in not obscuring the distinction, but even here the ringing answer of the headline (almost certainly not written by the journalist) is bolder and broader than the article claims.
A whimsical twitter thread of etymological onion peeling, now crystallised into a blog post, results in a splendid definition of AI. Starting with ‘complicated algorithms running on very fast computers’, we end up with AI helpfully described as
The method by which an old Persian magician uses counting stones, to move other stones, by way of amber resin, such that a casual observer thinks the stones are moving themselves.
In dealing with digital services – indeed in dealing with organisations generally – power is very asymmetric. Amazon does not invite you to negotiate the terms on which it is prepared to sell you things (though of course you retain the power not to buy). Digital services and apps give the illusion of control (let’s think about whether to accept these cookies…) but have developed a habit and a reputation or helping themselves to data and making their own judgement about what to do with it. That’s not necessarily because individual consumers can’t control permissions, but is also because the cost and complexity of doing so make it burdensome. Tom Steinberg brings a potential solution to that problem: what if we had somebody negotiating all that on our behalf, could that asymmetry be addressed? Typically, he recognises the difficulties as well as the potential, but even if the answers are hard, the question is important.
Erin Winick – MIT Technology Review
This devastatingly simple short post brings together estimates of the employment effects of automation, and assesses their consistency and coherence. There turns out to be none: ‘we have no idea how many jobs will actually be lost to the march of technological progress.’
A couple of weeks ago, the people of Hawaii were told that they were under missile attack. They weren’t, but that didn’t stop the warning being terrifying.
The cause was quickly narrowed down to poor user interface design. But poor user interface design is of course but one step in the chain of whys. This post follows several more links in the chain – giving a level of detail which at one level is more than most people will want or need, but using that to make some important points of much wider application. One is that critical designs need critical testing – and more generally that the value of design is not in the presence (or absence) of veneer. Another is that maintaining things is important and can be particularly difficult for systems funded on the basis that when they have been built, they are finished. The consequences of that approach may be irritating or they may be close to catastrophic, but they can be addressed only when there is recognition that, as David Eaves put it, you can either iterate before you fail, or you can do it after you fail, but you’ll do it either way.
Dave reads and reflects and shares both the reading and the reflections, on topics which are often closely linked to themes covered here. He has just announced a slightly different approach to sharing the material he finds, including a dedicated category on his blog (which comes with a selective RSS feed). Well worth following – though there is no obvious reason to filter out his own posts, which are always worth reading in their own right.
The easy mantra of ‘fail fast’ is one of many (mis)translated from agile thought and practice. The positive case is easy to understand, especially in contrast with slower project management approaches which consume all their time and money before discovering they have built the wrong thing. But in failing fast, the cost and the impact of the failure need to be understood too. In many public services that cost can be very high and, even more importantly, may fall on those least able to meet it.
This post is a powerful description of an extreme case of that – but in describing the extreme, there is plenty to reflect on for a much wider range of services. Sometimes failure is really not an option.