It was 3am and raining in the city by the Bay…

I’ve just read a review of Billy Wilder’s 1944 film noir classic Double Indemnity, wherein an insurance salesman is persuaded to help murder a woman’s husband and make it look like suicide. By selling the wife a policy that the husband doesn’t know about, with a ‘double indemnity’ clause increasing the payout if the insured person dies under very specific but unlikely circumstances – such as a fall from a moving train – the salesman sets up the wife’s plotting with a very profitable outcome. Of course, it all goes horribly wrong…

One of the supporting characters is Edward G. Robinson, who took a break from playing Chicago gangsters and played a claims manager in the salesman’s company with a nose for a fake claim. At one point in the film, he has this to say on the claim in question when justifying the time he’s spent on it to his boss. It has a bearing on the sort of mind-set that we as testers need:

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Come now, you’ve never read an actuarial table in your life, have you? Why they’ve got ten volumes on suicide alone. Suicide by race, by color, by occupation, by sex, by seasons of the year, by time of day. Suicide, how committed: by poison, by firearms, by drowning, by leaps. Suicide by poison, subdivided by types of poison, such as corrosive, irritant, systemic, gaseous, narcotic, alkaloid, protein, and so forth; suicide by leaps, subdivided by leaps from high places, under the wheels of trains, under the wheels of trucks, under the feet of horses, from steamboats. But, Mr. Norton, of all the cases on record, there’s not one single case of suicide by leap from the rear end of a moving train.

(…)

Desk job? Is that all you can see in it? Just a hard chair to park your pants on from 9 to 5? Just a pile of papers to shuffle around and five sharp pencils and a scratchpad to make figures on? Maybe a little doodling on the side? Well, that’s not the way I look at it, Walter. To me, a claims man is a surgeon. That desk is an operating table and those pencils are scalpels and bone-chisels. And those papers are not just forms and statistics and claims for compensation. They’re alive. They’re packed with drama, with twisted hopes and crooked dreams.

++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

I’ve yet to come across a software app that’s “packed … with twisted hopes and crooked dreams” but the rest of it rings very true.

Advertisements

No, Minister

There’s a lot of talk within the testing community at the moment about conferences, and in particular the idea of “pay to talk”. This takes two forms. One is where conference organisers offer potential sponsors a conference slot in return for financial support. The other – which is what I’m more concerned with in this post, is the practice of not meeting conference speakers’ legitimate expenses, either for travel, subsistence and accommodation, or indeed any reimbursement for their time and effort in putting together a presentation in the first place.
This phenonomen is not unknown elsewhere. Creatives – writers, artists, photographers and musicians – have gotten used to various people trying to get them to do work for no financial return – “it’ll be good exposure” is the usual excuse, or “We have no budget for this”. The funniest iteration of this happened to me, when I was working out my redundancy notice in my last company. The company was moving to a new office, and I was asked if I would be prepared to work a weekend on testing the server migration. “We have no budget, so we can’t actually pay you for this,” said my line manager, “but we can offer you time off in lieu.” I pointed out, surprisingly politely, that the company was about to give me more time off than I could comfortably cope with, and so I declined their generous offer. Apparantly a colleague later said to my manager “That was courageous of you”, using the Civil Service meaning of the word ‘courageous’; to be fair, that manager was under a three-line whip from the senior management. This was the same management who omitted to have anyone test the links in their social media announcement when the new office was opened, so their exhortation to “Click HERE for a virtual tour of our fabulous new office!” came to nothing when the link didn’t work. I have to say that I took perverse pleasure in posting “You’d think someone would have tested this before posting. Oh, no, I forgot – you SACKED all your testers!”

It’s one thing when it’s some promoter or client trying to get something for nothing; and there’s a lively discussion online about creatives pushing back against this (the wonderful Clients from Hell website is just the tip of this particular iceberg). But when the Government decides it wants something for nothing, that elevates the whole argument to a completely new level.

At the end of May, a new post appeared on the Ministry of Testing website*. It started:

The Home Office is hosting up to 200 people for TESTCON’18, a conference for UK Government Software Engineers engaged in testing and QA. This year in central London we’re looking for individuals and companies to come and host a session on Wednesday 10th of October 2018.

The theme of this year’s conference is Insight and Innovation and we’re looking for engaging sessions to bring in new ideas and best practice into our sector.

Slots can be anywhere from 5 minutes to an hour and we can be flexible with the session format. Unfortunately we aren’t able to offer monetary incentives but it is a great opportunity to present to an engaged audience in a building not open to the public….

There was some immediate pushback against the OP; I got some retaliation in to defend the OP immediately, because the OP wasn’t going to be the person responsible, it wasn’t their fault, and equally as a civil servant he wouldn’t be able to respond in public as robustly as I could or even respond to criticism. I posted:

Within the UK Government sector, you are looking here at a mindset where attendees from other Government Departments will have their expenses covered by their home department, and any political attendees will be operating within the “Westminster bubble” and so not incurring expenses anyway.

Sadly, speaking expenses have been something of a political issue over the past few years, mainly because of past indiscretions by some high-profile speakers. This makes Government departments reluctant to be seen to be offering “jollies”. The people who suffer are the independent and SME speakers, who cannot justify giving HMG a freebie, and HMG itself, of course, who do not get the benefit of the experience of people who have something helpful to say. In the end, it only serves to reinforce groupthink amongst managers in Government computing as they will ultimately only be listening to opinions from either within the Government estate or from big consultancies with an eye to getting or retaining high-profile contracts (no names, no pack drill, but I’m sure you can fill in the likely names as well as I can).

In my experience (30 years in the Civil Service, albeit at a minor level), testing is not prioritised by many managers with responsibility for IT – which is not the same as saying “IT managers”. Ultimate responsibility for IT will, in so many Government departments, rest with a senior administrator and not an IT person.

The decision to make this a “no budget” event will have been taken at a higher level than (the OP’s) pay grade, so don’t shoot the messenger.

However, I was as irritated by this as a lot of other people, so I did the only thing possible; I wrote to my Member of Parliament.
For those unfamiliar with UK Parliamentary protocol, the best way to challenge something that a UK Government department does, especially if it’s on a general point of policy or principle, is to ask your MP to write to the Minister concerned. It doesn’t guarantee that you’ll get action, or a change in a policy; but it does usually mean that your query will be answered by someone rather higher up in the organisation than if you wrote directly. If it so happens that your MP knows a Minister personally, then your query may be discussed – informally – on a one-to-one basis. Again, there is no correlation between the level of contact and any action.

My MP is Edward Argar, a fairly low-profile Conservative back-bencher who failed to be selected for the candidacy of a number of seats before being parachuted into my constituency, which is a safe Tory seat. He was appointed as a junior minister in the Ministry of Justice in June of this year.

I wrote:

Dear Mr. Argar,
I have recently seen a call for attendees to a conference organised by the Home Office, Testcon ’18, described as “…a conference for UK Government Software Engineers engaged in testing and QA.” It is due to be held on 10thOctober this year.
I am a professional software tester with over 20 years’ experience in both the Government and private sectors. Software testing is a growing area, but amongst some IT professionals and managers responsible for IT, it struggles against a perception of being either “merely checking”, a process that acts as a “gateway” to product release or something that is tacked onto the end of a software development process, inevitably resulting in limited time for testing. Testers may get blamed for bugs getting released into the final releases of software, even though no tester will ever suggest that they can guarantee 100% bug-free software; and they are not responsible for bugs being in the software in the first place. The results of poor testing and the lack of a more involved role for testers in the whole business process of creating, designing, developing and releasing software can be seen regularly in news stories about computer failures in both public and private sectors.
What I am concerned about with regards to the Testcon ’18 conference is that the call for e_xpressions of interest states that “Unfortunately we aren’t able to offer monetary incentives but it is a great opportunity to present to an engaged audience”. This is extremely concerning, both to me and to very many software testing practitioners. The industry as a whole is pushing back against a phenomenon which has been identified as “pay to speak”; situations where speakers attend conferences at their own (or more often, their companies’) expense is seen as seriously restricting the diversity of voices heard in testing conferences. Diversity in testing is not just a good idea; when applications are being rolled out into use in the Real World, they are going to be used by any number of potential customers, potentially world-wide. It follows from that, that applications need to be tested by a wide range of different testers with different life experiences. We have all come across applications where the systems work perfectly well in terms of the computer code properly executing its instructions, but the layout, workflow and assumptions that the designer made about the likely user have rendered the system, in some cases, unusable. I have personally experienced “the computer says no” moments that have made me write to CEOs of major companies in despair.
The reputation of Government IT systems is such that a major Department seeking the participation of speakers should be opening itself to the widest possible range of ideas. Instead, by not offering to meet even the most basic of expenses, the Home Office (as sponsoring Department for this conference) is excluding a wide range of individual practitioners and SMEs who are doing the most innovative work in developing both test techniques and innovative thinking on testing and its place in software development. Instead, the conference is most likely to be packed with attendees from other Government Departments (which is good) and from the big consultancy firms who can cover attendees’ expenses (not so good), treating the exercise as one of keeping their names in the frame for future contracting work. These are likely to be the same four consultancies who have been implicated in a range of recent bad news stories about accountancy, business and the existence of what outwardly seems like a cartel divvying up major private and public sector contracts between them whilst not actually doing a very good job, or who have been the key partners in past Government IT systems that have failed embarrassingly.
Your colleague, the new Home Secretary, has considerable experience in business and commerce, and indeed as a former Business Secretary will, one assumes, recognise the importance of SMEs in the economy. The SME sector in IT is one of the great repositories of knowledge and innovation in the British economy; but by taking a policy decision to not offer expenses for attendees and participants in this conference, the Home Office is cutting itself off from fresh and innovative thinking in the field of testing, and setting itself up to succumb to the same old groupthink that has plagued Government for too long.
I would appreciate it if you would raise this matter with Mr. Javid; I should be very interested to know his thoughts on the matter. I shall look forward to hearing from you and would be happy to meet to discuss this matter further.
Yours sincerely,

I heard nothing for two months – this is rather longer a time than usual, but MPs do have some other issue occupying their time at the moment (!). However, just when I was beginning to consider writing again, I received this:

Dear Mr Day,

Thank you for your email below – my apologies for omitting to respond, which I hope you will forgive. As you may be aware, I am no longer in the Home Office, so cannot speak for their approach to Conferences and speakers, however I did take the opportunity to highlight your concerns to a Home Office Minister. 

The Home Office has considerable internal testing/ software expertise which it draws on to ensure systems are robust and secure and, of course, this is augmented by external perspectives and input. My understanding is that, while such external contributions to conferences and other forums are both welcome and valuable to all concerned, including those speaking, it is not possible to pay for such appearances at all conferences, and I am afraid that I understand that the Home Office has no plans to pay all attendees who speak at their conferences. 

I appreciate that this response may be disappointing, but hope that it may nonetheless be helpful.

Kind regards, Edward.

Edward Argar MP

Member of Parliament for Charnwood

After a little consideration, I replied again:

Dear Mr Argar,

Thank you for your response.

As a professional tester, of course I am disappointed with the Home Office’s response. I am equally concerned as a taxpayer that the response shows no sign that the Home Office is fully committed to opening itself to a wide range of external perspectives, or realises quite what reaction its decision not to pay attending external speakers will provoke. I repeat that by not paying external speakers, the Home Office will almost certainly only hear perspectives from large companies or partnerships who are prepared to subsidise their personnel to speak as a means of keeping their names in consideration for future projects. Meanwhile, innovative speakers from the SME and freelance sectors, who cannot afford to give “freebies”, will not be heard. This will have a long-term effect on the quality of applications procured or developed.

I first became aware of the Home Office conference through a Brighton-based company, Ministry of Testing (which styles itself after the well-known nightclub rather than any other Government department!). MoT (as it is invariably known) is a training provider and conference organiser, and also promotes the wider testing community. I write as a satisfied member of that community and attendee at MoT-organised conferences, but I have no other direct connection with MoT. Nonetheless, they have played a key role in facilitating community discussion on paying to speak at conferences, and I copy my response to their forum.

If this subject interests you, I would be happy to meet and discuss the implications for government and industry on software testing generally and the role of Government-level conferences in particular.

Yours sincerely,
Robert Day

So what am I trying to achieve here? Well, a change in policy by one of the major departments of State would be nice, but I don’t see that happening under the present Government very much. A Government that expects people to work for nothing and has no shame in asking for that is frankly not fit for office; but saying so too loudly to an MP who you are trying to influence isn’t the best plan. (At least, not just yet.).

More specifically, though, I suppose I’m trying to raise the profile of testing as a whole, to get people in High Places to think a little bit more about testing and the role it plays in our lives. And perhaps if I could achieve that (though I’m not holding my breath), then the idea that testing as a professional discipline that should be taken notice of might just spread a bit wider than it does at the moment outside of the community. Someone has to try, and it might as well be me.

*For the benefit of anyone coming to this blog for the first time, “Ministry of Testing” is not a UK Government ministry. Rather, it is a software testing training provider and conference organiser, who just happen to style themselves after the famous London nightclub, Ministry of Sound.

The Computer says “You’re fired”.

There’s been some little kerfuffle online over this story from the wonderful world of corporate HR systems:

https://www.bbc.co.uk/news/technology-44561838

And here’s the guy’s story in his own words:

https://idiallo.com/blog/when-a-machine-fired-me

Although it’s an example from the USA, it has relevance here in the UK even though employment law is rather different on this side of the water. (This post is written specifically from a UK perspective.) The problem was perhaps not so much the procedure itself (though I shall return to that later); it was the fact that various people tried to unwind that procedure but the system wouldn’t permit it. And there was no good reason for the original supervisor not to renew the contract (though it does suggest an area where the company’s HR policies were not aligned with business needs).

The OP makes the point that what eventually happened was that he was dismissed and then eventually re-hired after the termination process was completed. Nonetheless, that really shouldn’t have happened under those circumstances; it left the OP out of pocket and also involved him in reputational damage. The system design was based on 1) an assumption that any decision to terminate is automatically correct and unchallengeable, and 2) in any termination process, there are a series of steps to be taken that have to be taken without any option for challenge at the operational level – i.e. the security team were under a three-line whip from the system to escort the OP off the premises, even though they had had conversations with other managers and knew that there was an issue with the process that the management team were trying to reverse.

This, of course, is a real-life illustration of the Milgram experiment, where any instruction from an authority figure is complied with, no matter how unreasonable. It’s funny how often that seems to be borne out in experience.

Someone commented to me that perhaps there was an underlying issue with the OP’s attitude and the possibility of their posting rude blogs online. Personally, I think Ibrahim Diallo’s blog was completely fair and justified, and at no point does he name the company where this occurred (though some who commented accurately identified them based on their own experience). Interestingly, the risk of disparaging blogs doesn’t seem to have been considered in system design as there was no provision made to generate a non-disclosure agreement on contractor exit!

The use of the words “fired” and “job” reflect common usage rather than strict legal definitions, though it also has some bearing on the long-running debate on the status of workers in the gig economy going all the way back (for us in the UK) to the Inland Revenue’s original implementation of IR35. The thing is that there is little indication that the process would be any different for a contracted employee. The omission of a renewal action by a disgruntled supervisor would never be an admissible reason for dismissal – it certainly wouldn’t come under the heading of “some other substantial reason” for dismissal which is the catch-all reason in UK employment law which is automatically assumed to be fair. And if such a system caused such a “cascade dismissal” of a contracted employee, it might render the company liable for action, not only in an Employment Tribunal, but also in a civil court for consequential loss for putting in place a system which actually circumvented accepted procedures (although the ACAS code of practice on discipline leading to dismissal isn’t statutory, I doubt a court would take a favourable view of a system that ignored established custom and practice in such a well-established legal area).

From a system testing point of view, this entire incident illustrates areas where testers should have had greater involvement at the design stage, in challenging issues such as “is it possible to back out of the process once it has been initiated?” or “Does this actually comply with employment law?” (and hence the underlying question, “What’s the worst that can happen to us – the company – if this goes badly wrong?”). There certainly seems to have been a distinct lack of risk analysis when this system was being defined and designed; and any concerns raised by testers that may have been raised seem to have been ignored. And that’s a valid challenge to make, irrespective of variations of legal practice in different jurisdictions.

Conducting testing

I was reading a blog post from @GregPaciga on the Modern Testing Principles being developed by Alan Page and Brent Jensen (http://gregorytesting.ca/2018/06/discussing-the-modern-testing-principles/). I saw Alan give a presentation at TestBash Brighton on ‘Modern Testing’ and found his arguments persuasive.

In his post, Gregory was describing a discussion about the seven principles of Modern Testing, and in particular reporting on issues raised in the discussion over the fifth and (particularly) the seventh principles. Part of that seventh principle caused a few raised eyebrows amongst a discussion group comprised of testers:

7. We expand testing abilities and knowhow across the team; understanding that this may reduce (or eliminate) the need for a dedicated testing specialist

Gregory said that he was unsure whether disagreement with the second part of that statement, about reducing/eliminating the need for dedicated test specialists, was a knee-jerk reaction or not. (I got the feeling that he was saying ‘knee-jerk reaction’ as if that were a Bad Thing.) He talks about reducing the need for a testing specialist without eliminating the need for the function itself.

Looking at this, I was reminded of the orchestral conductor André Previn. Many years ago (when he was principal conductor of the London Symphony Orchestra) he did some tv documentaries on the role of the conductor. In one, he set the orchestra playing a piece and then stepped away from the podium and sat down. The orchestra continued quite happily without him to the end of the piece. So why, asked Previn, have a conductor?

Well, it was a very standard piece of repertoire, something the orchestra had played many times before. The individual players were all talented and able musicians. So the conductor wasn’t completely necessary. But consider what might happen if the orchestra had to play an unfamiliar or new piece, or had to integrate a new member who did not know the way they worked and played together? Then all the assumptions that each orchestra member might make, the unstated consensus that the orchestra would collectively have, might be thrown into confusion. Then the conductor acts as a unifying force, setting the tempo, or exploring matters of interpretation or meaning in rehearsal, and holding the whole together in public performance.

I think this is the same as expanding testing knowhow across the entire development team. But that does not eliminate the need for a dedicated testing specialist. Rather, it changes the test specialist’s role. They become the person who directs testing; the person who thinks about testing as a discipline and steers individuals on their own testing journey; the person who helps mould the team into a single unit and, indeed, orchestrates the testing effort to a successful delivery. The tester also can act as a fresh pair of eyes in the planning and design stages of a project; just as the conductor does, they can consider matters of meaning or interpretation outside the cutting edge of the performance itself and communicate this back to the team in the development process (which you might think of as being like rehearsal).

And of course, this sees software development as being, like music, an art as well as a science.

I for one welcome our new robot overlords

AI isn’t going away; and certainly the testing blogosphere keeps returning to the subject. And not just the blogosphere: in today’s Daily Telegraph business section (14 May 2018), their business correspondent James Titcomb said “Tech giants need to build ethics into AI from the start”. Looking in particular at the well-publicised demonstration of Google Duplex, with its conversation between a caller and a call centre where one party was an AI system and one was human, and which certainly proved a challenge to the Turing Test, but also at various autonomous driving systems , he added “Every frontier technology now needs to be built with at least some level of paranoia: some person asking ‘How could this be abused?’ ”

As a science fiction fan, I’ve been thinking about artificial intelligences of one sort or another since I was a teenager; and in one form or another, there has been thinking on the whole subject since the 1940s, long before AI could become a physical reality. Even in the real world of physical things, I remember my father – at that stage in his life, a railway signalling system designer and implementation engineer – talking about “automation” in the 1960s. Later, when computers evolved into the first microprocessor-driven machines (first coming to my attention with the Tandy (UK)/Radio Shack (US) TRS-80 and the Commodore PET. both machines that shipped with processors with mind-bendingly huge memories of 4k (!) in their entry models), a book by Dr. Christopher Evans, The Mighty Micro and a BBC television documentary of the same name made serious political waves with predictions of a future both terrifying and rich with possibility. That’s the future we’re living in now.

In later years, I talked, exchanged letters, and went on drinking sessions with people whose stock-in-trade as imagineers saw them thinking about AI. A late friend of mine worked with Stanley Kubrick on his development of the film that Steven Spielberg finally finished as the eponymous A.I.: Artificial Intelligence, and I later talked with others who worked either with Kubrick or Spielberg on different aspects of that film.

OLYMPUS DIGITAL CAMERA
The soul of a new machine?

Fast-forward, then, to the beginning of the year and a Ministry of Testing Masterclass on the subject. In the wake of this, a thread was started in the MoT Club (Ethics in Machine Learning), and I’ve pointed a couple of bloggers to this in the past few days. With this reawakening of interest, I thought it was time to share the discussion with some more people.

(The other participants were Ministry of Testing test ninjas punkmik, andrewkelly2555, kimberley, ceedubsnz, jacks and vjkumran. Thanks to all for their input.

The discussion started by putting the question of what we should do as testers if we spotted a system under test doing, or heading towards doing, something unethical. Should we really expect AIs to have any ethics at all? It then emerged that AI systems are already being encountered in recruitment situations, and how they “learnt” from their initial ‘training’ dataset or parameters. Sifting according to a list of technical specifications, for instance, might reject good candidates where soft skills were specifically being looked for, but these might not be easy to quantify to the levels that the intelligent system demanded. And if the managers who were gatekeeping the recruitment process didn’t themselves fully understand the criteria they were selecting for (say, in a situation where the budget holder for the post being recruited for wasn’t themselves a tester), then having an intelligent system doing the sifting might just lead to the same flawed sift only being done more quickly and economically…

If we’re going to look at the ethics of using AI, we should be certain that our existing, human systems are equally ethically based for a starter.

This question has been exercising the minds of science fiction writers for nearly seventy years! Most people will point to Isaac Asimov and his Three Laws of Robotics . (Asimov himself claimed that there were enough loopholes in the Three Laws to keep him profitably selling stories for the next forty years…) The Will Smith film based on Asimov’s robot stories, I, robot, posed a perfect example of something unethical happening for ethical reasons. The Will Smith character had a major hangup about robots because he had been involved in an accident where two cars ended up sinking in a river. A robot went into the river to save the humans in the cars, based on the First Law. But the robot weighed up the likelihood of saving both humans and decided that was not possible. Instead, it prioritised saving one human based on best chance of survival and utility of the saved human to society. It saved the Will Smith character, a policeman, before attempting to save the other human, a child, despite Smith ordering the robot to save the child (Second Law trumped by First Law).

Interestingly, the film developed robot motivations to the point where they were prepared to restrict human freedoms because humans do things to themselves that are harmful. This reflected the work of another classic science fiction writer, Jack Williamson, whose “Humanoids” had only one directive: “To serve and obey, and to keep men from harm”. Taking this to its logical conclusion, the “Humanoids” ended up keeping the entire population under chemical lockdown for their own good.

So the answer to the question “Should we really expect AI’s to have any ethics at all?” looks uncomfortably as if that, given that for thirty or forty years we have been told that the role of business is wealth creation and ethics has no – or at best, a secondary – role to play in maximising shareholder value, the answer is sadly clear. Future “Laws of Robotics” may be closer to the alternative proposed by David Langford:

1. A robot will not harm authorised Government personnel but will terminate intruders with extreme prejudice.

2. A robot will obey the orders of authorised personnel except where such orders conflict with the Third Law.

3. A robot will guard its own existence with lethal antipersonnel weaponry, because a robot is bloody expensive.

So how can we even begin to test for ethics?

Fortunately, we aren’t in Asimov’s scenario, where his robots were multi-purpose machines that were expected to learn how to deal with new situations almost without limit. The systems we are likely to be testing, in the near term at least, will at least be designed to do a specific task, so that will help define the range of ethics that designers will have to build into systems and that we will have to test for.

What it will require will be for designers, analysts and testers to look to a different set of expert bodies as sources to build an understanding of the ethical issues that a system might have to incorporate. So an HR system would need input on equality issues as well as employment law; for these, I would first of all look (in the UK, at least) to the Equality & Human Rights Commission (EHRC) and/or to some of the trades unions, especially those active in the public sector who have addressed such things in trying to keep in line employers who are supposed to take ethical issues into account. For finance and accountancy issues, I’d be looking to take advice from banks and other investment bodies who have identified an ethical dimension to their work, such as (again in the UK) the Co-operative Bank.

I think this is an evolving area and a possible whole new field of expertise which will combine traditional IT skills and a range of soft skills that the IT profession hasn’t necessarily been noted for in the past. Otherwise, we could find ourselves in an “I for one welcome our new robot overlords” situation before we know it!

Others are thinking along these lines, as this article by David Weinberger shows.

A lot of this boils down to very basic questions. Who is responsible for unforeseen outcomes of a new tool? The old saying is that only a bad workman blames their tools, and this seems very pertinent. Ultimately, there is a human at the bottom of all AI learning algorithms. It would be just the same if a human was training an animal, say a dog, to undertake criminal acts such as theft (by retrieving someone else’s property) at a simple command. The dog could not be charged with the criminal act; it would be the owner, as trainer, that was responsible.

(This article on Geoffrey Hinton at the University of Toronto is an interesting read https://torontolife.com/tech/ai-superstars-google-facebook-apple-studied-guy/ )

What we should really be thinking about are complex situations which go beyond simple IF > THEN (action) decisions. It will always be the programmer who is responsible for how ethical we make our systems. I think the debate has to be about the failsafes that we build into our systems to enable AIs to spot unethical – or ethically ambiguous – situations and either apply ethical subroutines to decide the correct course of action or to stop and flag the situation to a human (who may or may not take an ethical decision, of course). And this does mean that as systems get more complex, the more they will have to be designed to spot ambiguous situations or ones where there are unforeseen circumstances.

To bring the discussion back to testing and testers, there’s the role for testers in the requirements gathering and system definition stage – trying to foresee the unforeseeable and design ethical safeguards into system behaviours and test them once code is written. And that’s going to need a different sort of skills set to the ones that are currently fashionable.