While the online world has turned into an ecosystem of bots (by which we mean computer scripts that automatically handle repetitive and mundane tasks), our knowledge of how these automated agents interact with each other is rather poor. But being automata without capacity for emotions, meaning-making, creativity, or sociality, we might expect bot interactions to be relatively predictable and uneventful.

In their PLOS ONE article “Even good bots fight: The case of Wikipedia“, Milena Tsvetkova, Ruth García-Gavilanes, Luciano Floridi, and Taha Yasseri analyze the interactions between bots that edit articles on Wikipedia. They track the extent to which bots undid each other’s edits over the period 2001–2010, model how pairs of bots interact over time, and identify different types of interaction outcomes. Although Wikipedia bots are intended to support the encyclopaedia — identifying and undoing vandalism, enforcing bans, checking spelling, creating inter-language links, importing content automatically, mining data, identifying copyright violations, greeting newcomers, etc. — the authors find they often undid each other’s edits, with these sterile “fights” sometimes continuing for years.

They suggest that even relatively “dumb” bots may give rise to complex interactions, carrying important implications for Artificial Intelligence research. Understanding these bot-bot interactions will be crucial for managing social media, providing adequate cyber-security, and designing autonomous vehicles (that don’t crash..).

We caught up with Taha Yasseri and Luciano Floridi to discuss the implications of the findings:

Ed.: Is there any particular difference between the way individual bots interact (and maybe get bogged down in conflict), and lines of vast and complex code interacting badly, or having unforeseen results (e.g. flash-crashes in automated trading): i.e. is this just (another) example of us not always being able to anticipate how code interacts in the wild?

Taha: There are similarities and differences. The most notable difference is that here bots are not competing. They all work based on same rules and more importantly to achieve the same goal that is to increase the quality of the encyclopedia. Considering these features, the rather antagonistic interactions between the bots come as a surprise.

Ed.: Wikipedia have said that they know about it, and that it’s a minor problem: but I suppose Wikipedia presents a nice, open, benevolent system to make a start on examining and understanding bot interactions. What other bot-systems are you aware of, or that you could have looked at?

Taha: In terms of content generating bots, Twitter bots have turned out to be very important in terms of online propaganda. The crawlers bots that collect information from social media or the web (such as personal information or email addresses) are also being heavily deployed. In fact we have come up with a first typology of the Internet bots based on their type of action and their intentions (benevolent vs malevolent), that is presented in the article.

Ed.: You’ve also done work on human collaborations (e.g. in the citizen science projects of the Zooniverse) — is there any work comparing human collaborations with bot collaborations — or even examining human-bot collaborations and interactions?

Taha: In the present work we do compare bot-bot interactions with human-human interactions to observe similarities and differences. The most striking difference is in the dynamics of negative interactions. While human conflicts heat up very quickly and then disappear after a while, bots undoing each others’ contribution comes as a steady flow which might persist over years. In the HUMANE project, we discuss the co-existence of humans and machines in the digital world from a theoretical point of view and there we discuss such ecosystems in details.

Ed.: Humans obviously interact badly, fairly often (despite being a social species) .. why should we be particularly worried about how bots interact with each other, given humans seem to expect and cope with social inefficiency, annoyances, conflict and break-down? Isn’t this just more of the same?

Luciano: The fact that bots can be as bad as humans is far from reassuring. The fact that this happens even when they are programmed to collaborate is more disconcerting than what happens among humans when these compete, or fight each other. Here are very elementary mechanisms that through simple interactions generate messy and conflictual outcomes. One may hope this is not evidence of what may happen when more complex systems and interactions are in question. The lesson I learnt from all this is that without rules or some kind of normative framework that promote collaboration, not even good mechanisms ensure a good outcome.

Read the full article: Tsvetkova M, Garcia-Gavilanes R, Floridi, L, Yasseri T (2017) Even good bots fight: The case of Wikipedia. PLoS ONE 12(2): e0171774. doi:10.1371/journal.pone.0171774

Taha Yasseri and Luciano Floridi were talking to blog editor David Sutcliffe.

In many sectors, automation is seen as a threat due to the potential for job losses. By contrast, automation is seen as an opportunity in healthcare, as a way to address pressures including staff shortages, increasing demand and workloads, reduced budget, skills shortages, and decreased consultation times. Automation may address these pressures in primary care, while also reconfiguring the work of staff roles and changing the patient-doctor relationship.

In the interview below, Matt Willis discusses a project, funded by The Health Foundation, which looks at opportunities and challenges to automation in NHS England general practice services. While the main goal of the project is to classify work tasks and then calculate the probability that each task will be automated, Matt is currently conducting ethnographic fieldwork in primary care sites to understand the work practices of surgery staff and clinicians.

Since the first automated pill counting machine was introduced in 1970 the role of the pharmacist has expanded to where they now perform more patient consultations, consult with primary care physicians, and require greater technical skill (including a Pharm.D degree). While this provides one clear way in which a medical profession has responded to automation, the research team is now looking at how automation will reconfigure other professions in primary care, and how it will shape its technical and digital infrastructures.

We caught up with Matt Willis to explore the implications of automation in primary care.

Ed.: One finding from an analysis by Frey and Osborne is that most healthcare occupations (that involve things like social intelligence, caring etc.) show a remarkably low probability for computerisation. But what sorts of things could be automated, despite that?

Matt: While providing care is the most important work that happens in primary care, there are many tasks that support that care. Many of those tasks are highly structured and repetitive, ideal things we can automate. There is an incredible amount of what I call “letter work” that occurs in primary care. It’s tasks like responding to requests for information from secondary care, an information request from a medical supplier, processing a trusted assessment, and so on.

There is also generating the letters that are sent to other parts of the NHS — and letters are also triaged at the beginning of each day depending on the urgency of the request. Medical coding is another task that can be automated as well as medication orders and renewal. All of these tasks require someone working with paper or digital text documents and gathering information according to a set of criteria. Often surgeries are overwhelmed with paperwork, so automation is a potential way to make a dent in the way information is processed.

Ed.: I suppose that the increasing digitisation of sensors and data capture (e.g. digital thermometers) and patient records actually helps in this: i.e. automation sounds like the obvious next step in an increasingly digital environment? But is it really as simple as that?

Matt: Well, it’s never as simple as you think it’s going to be. The commonality of data originating in a digital format usually does make data easier to work with, manipulate, analyze, and make actionable. Even when information is entirely digital there can be barriers of interoperability between systems. Automation could even be automating the use of data from one system to the next. There are also social and policy barriers to the use of digital data for automation. Think back to the recent care.data debacle that was supposed to centralize much of the NHS data from disparate silos.

Ed.: So will automation of these tasks be driven by government / within the NHS, or by industry / the market? i.e. is there already a market for automating aspects of healthcare?

Matt: Oh yes, I think it will be a variety of those forces you mention. There is already partial automation in many little ways all over NHS. Automation of messages and notifications, blood pressure cuffs, and other medical devices. Automation is not entirely new to healthcare. The pharmacist is an exemplar health profession to look at if we want to see how automation has changed the tasks of a profession for decades. Many of the electronic health record providers in the UK have different workflow automation features or let clinicians develop workflow efficiency protocols that may automate things in specific ways.

Ed.: You say that one of the bottlenecks to automating healthcare is lack of detailed knowledge of the sorts of tasks that could actually be automated. Is this what you’re working on now?

Matt: Absolutely. The data from labour statistics is self-reported and many of the occupations were lumped together meaning all receptionists in different sectors are just listed under receptionist. One early finding I have that I have been thinking about is how a receptionist in the healthcare sector is different in their information work than a receptionist’s counterpart in another sector. I see this with occupations across health, that there are unique features that differentiate health occupations from similar occupations. This begs the need to tease out those details in the data.

Additionally, we need to understand the use of technologies in primary care and what tasks those technologies perform. One of the most important links I am trying to understand is that between the tasks of people and the tasks of technologies. I am working on not only understanding the opportunities and challenges of automation in primary care but also what are the precursors that exist that may support the implementation of automation.

Ed.: When I started in journals publishing I went to the post room every day to mail out hardcopy proofs to authors. Now everything I do is electronic. I’m not really aware of when the shift happened, or what I do with the time freed up (blog, I suppose..). Do you think it will be similarly difficult in healthcare to pin-point a moment when “things got automated”?

Matt: Well, often times with technology and the change of social practices it’s rarely something that happens overnight. You probably started to gradually send out less and less paper manuscripts over a period of time. It’s the frog sitting in a pot where the heat is slowly turned up. There is a theory that technological change comes in swarm patterns — meaning it’s not one technological change that upends everything, but the advent of numerous technologies that start to create big change.

For example, one of the many reasons that the application of automation technologies is increasing is the swarming of prior technologies like “big data” sets, advances in machine vision, machine learning, machine pattern recognition, mobile robotics, the proliferation of sensors, and further development of autonomous technologies. These kinds of things drive big advances forward.

Ed.: I don’t know if people in the publishing house I worked in lost their jobs when things like post rooms and tea trolleys got replaced by email and coffee machines — or were simply moved to different types of jobs. Do you think people will “lose their jobs“ as automation spreads through the health sector, or will it just drive a shift to people doing something else instead?

Matt: One of the justifications in the project is that in many sectors automation is seen as a threat, however, automation is seen as an opportunity in healthcare. This is in great part due to the current state of the NHS and that the smart and appropriate application of automation technologies can be a force multiplier, particularly in primary care.

I see it as not that people will be put out of jobs, but that you’ll be less likely to have to work 12 hours when you should be working 8 and to not have a pile of documents stacking up that you are three months behind in processing. The demand for healthcare is increasing, the population is aging, and people live longer. One of the ways to keep up with this trend is to implement automation technologies that support healthcare workers and management.

I think we are a long ways away from the science fiction future where a patient lays in an entirely automated medical pod that scans them and administers whatever drug, treatment, procedure, or surgery they need. A person’s tasks and the allocation of work will shift in part due to technology. But that has been happening for decades. There is also a longstanding debate about if technology creates more jobs in the long term than it destroys. It’s likely that in healthcare we will see new occupational roles, job titles, and tasks emerge that are in part automation related. Also, that tasks like filing paperwork or writing a letter will seem barbaric when a computer can, through little time and effort, do that for you.

Matthew Willis was talking to blog editor David Sutcliffe.