How ethical is your ethical robot?

If you're in the business of making ethical robots, then sooner or later you have to face the question: how ethical is your ethical robot? If you've read my previous blog posts then you will probably have come to the conclusion 'not very' - and you would be right - but here I want to explore the question in a little more depth.

First let us consider whether our 'Asimovian' robot can be considered ethical at all. For the answer I'm indebted to philosopher Dr Rebecca Reilly-Cooper who read our paper and concluded that yes, we can legitimately describe our robot as ethical, at least in a limited sense. She explained that the robot implements consequentialist ethics. Rebbeca wrote:

"The obvious point that any moral philosopher is going to make is that you are assuming that an essentially consequentialist approach to ethics is the correct one. My personal view, and I would guess the view of most moral philosophers, is that any plausible moral theory is going to have to pay at least some attention to the consequences of an action in assessing its rightness, even if it doesn’t claim that consequences are all that matter, or that rightness is entirely instantiated in consequences. So on the assumption that consequences have at least some significance in our moral deliberations, you can claim that your robot is capable of attending to one kind of moral consideration, even if you don’t make the much stronger claim that is capable of choosing the right action all things considered."

One of the great things about consequences is that they can be estimated - in our case using a simulation-based internal model which we call a consequence engine. So from a practical point of view it seems that we can build a robot with consequentialist ethics, whereas it is much harder to think about how to build a robot with say Deontic ethics, or Virtue ethics.

Having established what kind of ethics that our ethical robot has, now consider the question of how far does the robot go toward moral agency. Here we can turn to an excellent paper by James Moor, called The Nature, Importance and Difficulty of Machine Ethics. In that paper* Moor suggests four categories of ethical agency - starting with the lowest. Let me summarise those here:

1. Ethical impact agents: Any machine that can be evaluated for its ethical consequences.
2. Implicit ethical agents: Designed to avoid negative ethical effects.
3. Explicit ethical agents: Machines that can reason about ethics.
4. Full ethical agents: Machines that can make explicit moral judgments and justify them.

The first category: ethical impact agents, really includes all machines. A good example is a knife, which can clearly be used for good (chopping food, or surgery) or ill (as a lethal weapon). Now think about the blunt plastic knife that comes with airplane food - that falls into Moor's second category since it has been designed to reduce the potential of ethical misuse - it is an implicit ethical agent. Most robots fall into the first category: they are ethical impact agents, and a subset - those that have been designed to avoid harm by, for instance detecting if a human walks in front of them and automatically coming to a stop - are implicit ethical agents.

Let's now skip to Moor's fourth category, because it helps to frame our question - how ethical is your ethical robot? At present I would say there are no machines that are full ethical agents. In fact the only full ethical agents we know are 'adult humans of sound mind'. The point is this - to be a full ethical agent you need to be able to not only make moral judgements but account for why you made the choices you did.

It is clear that our simple Asimovian robot is not a full ethical agent. It cannot choose how to behave (like you or I), but is compelled to make decisions based on the harm-minimisation rules hard-coded into it. And it cannot justify those decisions post-hoc. It is, as I've suggested elsewhere, an ethical zombie. I would however argue that because of the cognitive machinery the robot uses to simulate ahead to model and evaluate the consequences of each of its next possible actions combined with its safety/ethical logical rules to choose between those actions, then the robot can be said to be reasoning about ethics. I believe our robot is an explicit ethical agent in Moor's scheme.

Assuming you agree with me, then does the fact that we have reached the third category in Moor's scheme mean that full ethical agents are on the horizon? The answer is a big NO. The scale of Moor's scheme is not linear. It's a relative small step from ethical impact agents to implicit ethical agents. Then a very much bigger step to explicit ethical agents, which we are only just beginning to take. But there is a huge gulf then to full ethical agents, since they would almost certainly need something approaching human equivalent intelligence.

But maybe it's just as well. The societal implications of full ethical agents, if and when they exist, would be huge. For now at least, I think I prefer my ethical robots to be zombies.

---

*Moor JH (2006), The Nature, Importance and Difficulty of Machine Ethics, IEEE Intelligent Systems, 21 (4), 18-21.

My contribution to an Oral History of Robotics

In March 2013 I was interviewed by Peter Asaro for the IEEE.tv Oral History of Robotics series. That interview has now been published, and here it is:

In case you're wondering, I'm sitting in Noel Sharkey's study (shivering slightly - it was a bitterly cold day in Sheffield). It was a real privilege to be asked to contribute, especially alongside the properly famous roboticists Peter interviewed. Do check them out. There doesn't seem to be an index page, but the set starts on page 2 of the IEEE.tv history channel.

Postscript: a full transcript of the interview can be found here: http://ethw.org/Oral-History:Alan_Winfield

Towards ethical robots: an update

This post is just a quick update on our ethical robots research.

Our initial ethical robot experiments were done with e-puck robots. Great robots but not ideal for what is essentially human-robot interaction (HRI) research. Thus we've switched to NAO robots and have spent the last few months re-coding for the NAOs. This is not a trivial exercise. The e-pucks and NAO robots have almost nothing in common, and colleague and project post-doc Dieter Vanderelst has re-created the whole consequence engine architecture from the ground up, together with the tools for running experiments and collecting data.

Why the NAO robots? Well, they are humanoid and therefore better fitted for HRI research. But more importantly they're much more complex and expressive than the e-puck robots, and provide huge scope for interesting behaviours and responses, such as speech or gesture.


But we are not yet making use of that additional expressiveness. In initial trials Dieter has coded the ethical robot to physically intervene, i.e. block the path of the 'human' it order to prevent it from coming to harm. Here below are two example runs, illustrated with composite images showing overlaid successive screen grabs from the overhead camera.

Here red is the ethical robot, which is initially heading toward its goal position toward the top right of the arena. Meanwhile blue - the proxy human - is not looking where it's going and heading for danger, at the bottom right of the arena. When it notices this red then diverts from its path, blocks blue, which then simply halts. Red's consequence engine now predicts no danger to blue and so red resumes progress toward its goal.

Although we have no videos just yet you can catch a few seconds of Dieter and the robots at 2:20 on this excellent video made by Maggie Philbin and Teentech for the launch of the BBC Micro:bit earlier this month.

Copyright violations bring Forth memories

Last week I found a couple of copyright violations. Am I upset? Not at all - actually I'm delighted that stuff I did in 1983 is alive and well on the Interweb thanks to the efforts of others.

The first is an online readable copy of my 1983 book The Complete Forth. It's a textbook on the programming language Forth that I was heavily into at the time. The book was first published by Sigma Press, then internationally by John Wiley, and was translated into both Dutch and Japanese. Someone - I assume from the Jupiter Ace archive - has gone to the trouble of scanning every page. Even the pull out reference card. Whoever you are, thank you.

Just before I wrote that book, I had developed a Forth programming system (a programming environment that integrates compiler and interpreter) for the NASCOM 2 Z80 micro computer. Myself and a friend then marketed Hull-Forth and, I recall, sold several hundred copies. Of course this was pre-internet so marketing meant small ads in the magazine Personal Computer World. What we actually shipped was a printed manual together with the code on a cassette tape. Floppy disks were hugely expensive and beyond the reach of hobby computers, so for saving and loading programs we used audio cassette recorders. They were slow and very unreliable; if there was a checksum error you just had to rewind, cross your fingers and try again. I can't imagine anyone feeling nostalgic for that particular technology.

This brings me to the second copyright infringement. By accident I discovered there is a webpage for NASCOM enthusiasts, and several emulators, so you can run a virtual NASCOM on your modern PC. Scrolling down the long list of software in the NASCOM repository, in the section Programming Languages, I find

• Hull Forth Interpreter/Compiler (1982)

Hah! Someone must have gone to alot of trouble to get the code from the original cassette*, recorded using the Kansas City Standard at, I think, 300 baud (so slow you could almost hear the noughts and ones!), to a .NAS file you can download into your NASCOM emulator.

Ok, now to get that NASCOM emulator running. It will be fun (and slightly absurd) to run Hull Forth again for the first time in about 33 years.

---

*I probably still have one of those cassettes in my loft**, but no way of reading the data from it.
**Along with stacks of punched cards, rolls of paper tape, and all kinds floppy disks.

Forgetting may be important to cultural evolution

Our latest paper from the Artificial Culture project has just been published: On the Evolution of Behaviors through Embodied Imitation.

Here is the abstract

This article describes research in which embodied imitation and behavioral adaptation are investigated in collective robotics. We model social learning in artificial agents with real robots. The robots are able to observe and learn each others' movement patterns using their on-board sensors only, so that imitation is embodied. We show that the variations that arise from embodiment allow certain behaviors that are better adapted to the process of imitation to emerge and evolve during multiple cycles of imitation. As these behaviors are more robust to uncertainties in the real robots' sensors and actuators, they can be learned by other members of the collective with higher fidelity. Three different types of learned-behavior memory have been experimentally tested to investigate the effect of memory capacity on the evolution of movement patterns, and results show that as the movement patterns evolve through multiple cycles of imitation, selection, and variation, the robots are able to, in a sense, agree on the structure of the behaviors that are imitated.

Let me explain.

In the artificial culture project we implemented social learning in a group of robots. Robots were programmed to learn from each other, by imitation. Imitation was strictly embodied, so robots observed each other using their onboard sensors and, on the basis of only visual sense data from a robot’s own camera and perspective, the learner robot inferred another robot’s physical behaviour. (Here is a quick 5 minute intro to the project.)

Not surprisingly embodied robot-robot imitation is imperfect. A combination of factors including the robots’ relatively low-resolution onboard camera, variations in lighting, small differences between robots, multiple robots sometimes appearing within a learner robot’s field of view, and of course having to infer a robot’s movements by tracking the relative size and position of that robot in the learner’s field of view, lead to imitation errors. And some movement patterns are easier to imitate than others (think of how much easier it is to learn the steps of a slow waltz than the samba by watching your dance teacher). The fidelity of embodied imitation for robots, just as for animals, is a complex function of four factors: (1) the behaviours being learned, (2) the robots’ sensorium and morphology, (3) environmental noise and (4) the inferential learning algorithm.

But rather than being a problem, noisy social learning was our aim. We are interested in the dynamics of social learning, and in particular the way that behaviours evolve as they propagate through the group. Noisy social learning means that behaviours are subject to variation as they are copied from one robot to another. Multiple cycles of imitation (robot B learns behaviour m from A, then robot C learns the same behaviour m′ (m mutated), from robot B, and so on), gives rise to behavioural heredity. And if robots are able to select which learned behaviours to enact we have the three Darwinian operators for evolution, except that this is behavioural, or memetic, evolution.

These experiments show that embodied behavioural evolution really does take place. If selection is random, that is robots select which behaviour to enact from those already learned – with equal probability, then we see several interesting findings.

1. If by chance one or more high fidelity copies follow a poor fidelity imitation, the large variation in the initial noisy learning can lead to a new behavioural species, or traditions. Thus showing that noisy social learning can play a role in the emergence of novelty in behavioural (i.e. cultural) evolution. That was written up in Winfield and Erbas, 2011.

But it is the second and third findings that we describe in our new paper.

2. We see that behaviours appear to adapt to be easier to learn, i.e. better ‘fitted’ to the robot swarm. The way to think about this is that the robots' sensors and bodies, and physical environment of the arena with several robots (including lighting), together comprise the 'ecological niche' for behavioural evolution. Behaviours mutate but the ones better fitted to that niche survive.

3. The third finding from this series of experiments is perhaps the most unexpected and the one I want to outline in a bit more detail here. We ran the same embodied behavioural evolution with three memory sizes: no memory, limited memory and unlimited memory.

In the unlimited memory trials each robot saved every learned meme, so the meme pool across the whole robot population (of four robots) grew as the trial progressed. Thus all learned memes were available to be selected for enaction. In the limited memory trials each robot had a memory capacity of only five learned memes, so that when a new meme was learned the oldest one in the robot's memory was deleted.

The diagram below shows the complete family tree of evolved memes, for one typical run of the limited memory case. At the start of the run the robots were seeded with two memes, shown as 1 and 2 at the top of the diagram. Behaviour 1 was a movement pattern in which the robot traces a triangle path, behaviour 2 a square. Because this was a limited memory trial the total meme pool has only 20 behaviours - these are shown below as diamonds. Notice the cluster of 11 closely related memes at the bottom right, all of which are 7th, 8th or 9th generation descendents of the triangle meme.

Behavioural evolution map following a 4-robot experiment with limited memory; each robot stores only the most recent 5 learned behaviours. Each behaviour is descended from two seed behaviours labelled 1 and 2. Orange nodes are high fidelity copies, blue nodes are low fidelity copies. The 20 behaviours in the memory of all 4 robots at the end of the experiment are highlighted as diamonds. Note the cluster of 11 closely-related behaviours at the bottom right.

When we ran multiple trials of the limited and unlimited memory cases, then analysed the number and sizes of the clusters of related memes in the meme pool, we saw that the limited memory trials showed a smaller number of larger clusters than the unlimited memory case. The difference was clear and significant; with limited memory an average of 2.8 clusters of average size 8.3, with unlimited memory 3.9 clusters of size 6.9.

Why is this clustering interesting? Well it's because the number and size of clusters in the meme pool are good indicators of its diversity. Think of each cluster of related memes as a 'tradition'. A healthy culture needs a balance between stability and diversity. Neither too much stability, i.e. a very small number (in the limit 1) of traditions, or too much diversity, i.e. clusters so small that there are no persistent traditions at all. Perhaps the ideal balance is a smallish number of somewhat persistent traditions.

So far I didn't mention the no memory case. This was the least interesting of the three. Actually by no memory we mean a memory size of one; in other words a robot has no choice but to enact the last behaviour it learned. There is no selection, and no clusters can form. Traditions can never even get started, let alone persist.

Of course it would unwise to draw any big conclusions from this limited experimental study. But an intriguing possibility is that some forgetting (but not too much) may, just like noisy imitation, be a necessary condition for the emergence of culture in social agents.

Full reference:
Erbas MD, Bull L and Winfield AFT (2015), On the Evolution of Behaviors through Embodied Imitation, Artificial Life, 21 (2), pp 141-165. The full text (final draft) paper can be downloaded here.

Related blog posts:
Robot imitation as a method for modelling the foundations of social life
Open-ended Memetic Evolution, or is it?

Yesterday I looked through the eyes of a robot

It was a NAO robot fitted with a 3D printed set of goggles, so that the robot has two real cameras on its head (the eyes of the NAO robot are not in fact cameras). I was in another room wearing an Oculus Rift headset. The Oculus was hooked up to the NAO's goggle cameras, so that I could see through those cameras - in stereo vision.

photo by Peter Gibbons

But it was even better than that. The head positioning system of the Oculus headset was also hooked up to the robot, so I could turn my head and - in sync - the robot's head moved. And I was standing in front of a Microsoft Kinect that was tracking my arm movements. Those movements were being sent to the NAO, so by moving my arms I was also moving the robot's arms.

All together this made for a pretty compelling immersive experience. I was able look down while moving my (robot) arms and see them pretty much where you would expect to see your own arms. The illusion was further strengthened when Peter placed a mirror in front of the NAO robot, so I could see my (robot) self moving in the mirror. Then it got weird. Peter asked me to open my hand and placed a paper cup into it. I instinctively looked down and was momentarily shocked not to see the cup in my hand. That made me realise how quickly - within a couple of minutes of donning the headset - I was adjusting to the new robot me.

This setup, developed here in the BRL by my colleagues Paul Bremner and Pete Gibbons, is part of a large EPSRC project called Being There. Paul and Peter are investigating the very interesting question of how humans interact with and via teleoperated robots, in shared social spaces. I think teleoperated robot avatars will be hugely important in the near future - more so than fully autonomous robots. But our robot surrogates will not look like a younger buffer Bruce Willis. They will look like robots. How will we interact with these surrogate robots - robots with human intelligences, human personalities - seemingly with human souls? Will they be treated with the same level of respect that would be accorded their humans if they were actually there, in the flesh? Or be despised as voyeuristic; an uninvited webcam on wheels?

Here is a YouTube video of an earlier version of this setup, without the goggles or Kinect:

I didn't get the feeling of what it is like to be a robot, but it's a step in that direction.

Like doing brain surgery on robots

I spent a rare few hours in the lab the last few days, actually doing research. Or at least attempting to. Actually I made no progress at all. But did reach base camp: I managed to set up and run the ethical-dilemma robot experiment. And in the process refreshed my rusty command-line Linux. I was also reminded how time consuming, and downright frustrating experimental robotics research really is. Here's a taste: everything is set up and looks ok... but wait - the tracking system needs recalibrating; hmm... where's the manual? Ah, found it. Ok, wow this is complicated. Needs the special calibrating wand, and set square device... An hour later: ok ready now. Start everything up. But one of the robots isn't connecting. Ah, battery low, ok battery changed, now back up 4 steps and restart. And so it goes.

This is Swarmlab mission control. Three computers, three different operating systems;) The one in the middle (Windows XP) is running the Vicon tracking system, and monitoring via an overhead webcam. The laptop on the left (Ubuntu Linux) is running the four different processes to start and manage the three robots.

Here are the three e-pucks, each a WiFi networked Linux computer (Debian) in its own right. Actually each robot has two processors: a low-level PIC microcontroller to take care of motor control, managing the robot's sensors, etc. And an ARM processor for high-level control. The two interfaced via the SPI bus.







The setup is complicated. 5 computers in total, running a total of 9 networked processes. Here's a diagram showing those processes and how they are linked.

So, back to research.

The task I had set myself was to make some small changes to the high level controller. How hard can that be, you might think? Well it feels a bit like brain surgery: trying to tease apart code that I barely understand without breaking it. The code is well written and well structured, but it's in Python, which is new to me. It's only a couple of hundred lines, but like the neo-cortex - it's a thin layer at the top of a complex network of carefully choreographed processes and subsystems.

---

Acknowledgements: Christian Blum programmed the ethical robot experiments, supported by Dr Wenguo Liu who designed and setup the Swarmlab experimental infrastructure, including the e-puck Linux extension boards.

Surgical micro-robot swarms: science fiction, or realistic prospect?

Imagine a swarm of microscopic robots that we inject into the vascular system: the swarm swims to the source of the problem, then either delivers therapeutics or undertakes microsurgery directly.

That was how I opened a short invited talk at the Royal Society of Medicine on 5 February, at a meeting themed The Future of Robotics in Surgery. The talk was a wonderful opportunity for me to introduce swarm intelligence and speculate on the likelihood of surgical micro-robot swarms, while at the same time learning about robot surgery. Here are the slides from my talk (with links to YouTube videos where available).



The talk was in three parts.

First I introduced swarm intelligence, and its artificial counterpart swarm robotics. I showed, with examples from two of my students, how - with very simple rules - a swarm of robots can keep together as a swarm, while moving toward a beacon. Then, with a phagocyte-like behaviour, encapsulate the beacon. In our case these were lab robots moving toward an infra-red beacon, but it's not hard to imagine the same behavioural rules in a microscopic swarm swimming toward the source of a chemical marker (chemotaxis). I then gave two examples of the state of the art in swarm robotics: SYMBRION and (my current favourite) TERMES. I wanted to illustrate emergent physical interaction, in these two cases swarm self-assembly and swarm construction, respectively.

In part two I outlined what is by far the biggest problem: actually engineering robots at the micro-scale. Here I drew upon the examples from my book Robotics: a very short introduction; a section called A swarm of medical microrobots.  Start with cm sized robots. These already exist in the form of pillbots and I reference the work of Paolo Dario's lab in this direction. Then get 10 times smaller to mm sized robots. Here we're at the limit of making robots with conventional mechatronics. The almost successful I-SWARM project prototyped remarkable robots measuring 4 x 4 x 3mm. But now shrink by another 3 orders of magnitude to microbots, measured in micrometers. This is how small robots would have to be in order to swim through and access (most of) the vascular system. Here we are far beyond conventional materials and electronics, but amazingly work is going on to control bacteria. In the example I give from the lab of Sylvain Martel, swarms of magnetotactic bacteria are steered by an external magnetic field and, interestingly, tracked in an MRI scanner.

In the final part of my talk I introduce the work of my colleague Sabine Hauert, on swarms of nanoparticles for cancer nanomedicine. These 5 - 500nm particles are controlled by changing their body size, material, coating and cargo so - in true swarm fashion - the way the nanoparticle swarm moves and interacts with much larger normal and tumour cells is an emergent property of the way the nanoparticles individually interact and cooperate. Sabine and her collaborators have created an online tool called NanoDoc, which allows anyone to edit the design of nanoparticles then run simulations to see how their designs perform. In this way the task of searching the huge design space is crowd-sourced. In parallel Sabine is also running mesoscale embodied simulations, using the Harvard Kilobots.

I concluded by suggesting that engineering micro or nanobots is not the only major challenge. At least as important are: (a) how would you program the swarm, and (b) how would such a swarm be approved for clinical use? But a deeply interesting question is the nature of the human-swarm interface. If a swarm of surgical microbots should become a practical proposition would we treat the swarm as a microscopic instrument under the surgeon’s control, or a smart drug that does surgery?

Maybe we need an Automation Tax

Imagine this situation. A large company decides to significantly increase the level of automation at one of its facilities. A facility that currently employs a substantial number of men and women doing relatively low-skill tasks, which can now be done by a new generation of robots. Most of the workers get laid off which, for them and their families, leads to real hardship. The company was the only large employer in the area, which is economically depressed (one of the reasons perhaps that the company built the facility there in the first place), so finding alternative work is really difficult. And because most of those jobs were minimum wage, with little or no job security, redundancy payouts are small or non-existent and this of course means that the laid-off workers have no financial buffer to help them re-skill or relocate.

Now I am not anti-automation. Absolutely not. But I believe very strongly that the benefits of robotics and automation should be shared by all. And not just the shareholders of a relatively small number of very large companies. After all, the technology that such companies benefit from was developed by publicly funded research in university research labs. In other words research that you and I funded through our taxes. Ok, you might say, but companies pay tax too, aren't those taxes also contributing to that research? Yes, that's true. But large companies are very good at reducing their tax bill, multinationals especially. Our imaginary company may, in reality, pay most of its tax in a different country entirely from the one hosting the facility.

And of course it is we, through local and national taxation, who - as best we can - pick up the pieces to support the laid off workers and their families, through family tax credits, employment and support allowance, and so on.

Maybe we need an Automation Tax?

It would be a tax levied nationally, whenever a company introduces robotics and automation to one of its facilities in that country. But the tax would only be payable under certain conditions, so for instance:

• If the new robotics and automation causes no-one to be laid off, then no tax is due.

• If the automation does result in jobs becoming redundant, but the company re-trains and re-deploys those workers within its organisation, then no tax is due.

• If the company does lay off workers but makes a tax-free redundancy payment to those workers - regardless of their contract status - sufficient to cover the full costs of retraining, upskilling, and - with all reasonable efforts - finding work elsewhere, then no tax is due.

Only if none of these conditions are met, would the automation tax be due. The idea is not to discourage automation, but to encourage companies to accept a high degree of responsibility to workers laid off as a result of automation, and more widely their social responsibility to the communities in which they are located. The tax would enforce the social contract between companies and society.

Of course this automation tax doesn't go anywhere near far enough. I think the best way of sharing the wealth created by robotics and automation is through a universal Basic Income, but until that utopian condition can be reached, perhaps an automation tax is a start.