Science needs hacking: to keep it open, mutating and evolving
Esvelt's article in Nature magazine - complete with my highlights. Click to enlarge.
Have you heard of gene-drive elements before? Most likely not. This hugely powerful genetic engineering technology is currently being tested out and pioneered in the wild in Brazil by company Oxitec. The potential impact of this technology on wild populations of mosquitoes is huge. Yet, most likely, you haven't heard about it. And that means, most likely people in charge of regulations and other forms of governance haven't heard about it either.
Kevin Esvelt is group leader of the MIT Media Lab "Sculpting Evolution" group and pioneers some amazing research. Gene-drive systems are a new technology within genetic engineering and synthetic biology which could have a huge impact on our natural mosquito populations, and thus malaria frequency and threat. What's really interesting about this technology, is just how fast the turnaround of the innovation and research has been, while understanding and regulation by official bodies is lagging behind. See here's the deal: two years ago, research was published around the technological innovation in genetic engineering which would allow gene-drive elements to be implemented in multiple organisms. Now, two years on, guidelines are released for the responsible conduct of this gene-drive research by the US National Academy of Sciences finally released. The problem is: in that time, gene drive systems were already implemented in four species - before regulations outlined what was considered safe or not.
I first heard about gene-drive elements during a talk at Imperial College, by a group using genetic engineering to control populations of malaria-transmitting mosquitoes. To date, all successful malaria control programmes have aimed at reducing the population of the mosquito vector, either by habitat removal or insecticide spraying. Instead, what the group have managed to do, is to create a genetic element which selectively disrupts female fertility, reducing the reproductive output and leading to a population crash. Obviously, this kind of gene would not be favoured by evolution and would, if natural systems were at play, be diluted pretty quickly out of the population. This is where gene drive comes in. A gene drive element is one that is biased. The gene has some sort of characteristic which means that it is inherited by a disproportionately high proportion of the next generation - even if that gene has disastrous consequences for the organism. Depending on the fine-tuning of the gene drive element, this can mean that an introduced, artificial gene can very very quickly spread throughout a population of organisms and quickly become present in a large proportion of the next generation.
The reason this is exciting for scientists, is that if the aim is to genetically engineer a whole population of organisms, e.g. in the wild, it used to be near impossible to achieve the spread of your artificially introduced gene, without engineering a very large proportion of the population. Now, with gene drive, a gene can be engineered so that with just a few introductions into the population, scientists can sit back, allow reproduction to take place, and be safe in the knowledge that their gene will always be disproportionately likely to make it to the next generation.
But, as with all powerful technology, the opportunity for the power of the technology to lead to unwanted impact and effects is cause for concern. The very nature of this technology is that it has a run-away, accelerating spread throughout a population. It is uncontrollable in nature. If a gene-drive element finds its way into natural populations, without meaning to, it may already become too late to work out the far-reaching consequences to the wider ecological system before the majority of the population is carrying that gene.
I remember asking a provocative question at the end of the talk on mosquito gene drive: asking whether the researchers had considered the impact on the ecosystem and environment of releasing this promiscuous and potentially lethal genetic technology into the natural environment. There was very little in the way of solid answers from the researchers presenting their work, and the atmosphere in the room was awkward and tense. This is cause for concern: the problem isn't that we are becoming unethical in nature, its that the power and impact of our technology is swelling to the point where mistakes or short-sightedness will have larger and larger consequences.
This is why Esvelt's commentary on the closed-door nature of science is perfectly on point, and timely. I often like to think about what a system would look like if we invented it right now, on the spot, instead of accepting the handed-down, iterated, mutated versions that we find ourselves carrying out because it's the way that it's done. If we invented the scientific enterprise again, it would be unrecognisable from the current system. Academic science, as it stands, is as if it was designed by dinosaurs. For dinosaurs. It's burocratic, encourages anti-collaboration, results in frustration, inefficiency and stunts creativity, innovation, while cramping the inherently dynamic style of research. It's also driven by money, which instead of funding a person or great idea, often comes in the form of grants, which take weeks to write, months to apply for, and up to years to arrive. It's insane.
Esvelt's article got my cogs whirring, and a lot of the old frustrations and ideas I experienced when first starting in this field of academia again resurfaced. Science needs innovation. It's just that part of what needs innovating is the ability to innovate, collaborate and understand the problems at all -- i.e. the closed-door nature of science inhibits the innovation of its closed door nature.
So, here are some ideas from someone who stands both out and in of this closed door - I often feel like a bridge between the stuffy, hugely privileged and somewhat isolated academic institutions and between laymen, people on the ground, hackers, thinkers and activists.
Friends in and around science and technology, let's hack these problems:
1. Labs and academic groups often have no idea what other people in their building, let alone the whole world, are working on.
How can we break down these walls between groups? How can we set up a structure for fair, open, transparent collaboration? How can we change the culture of scientific research?
I am already developing an idea which allows easy, within-University collaborations to happen more easily, and be recognised/rewarded. I believe providing the tools and systems within these institutions will play a major role in cajoling researchers to innovate on their understanding of how to do research.
It's so drilled into academic scientists that they have to make their discovery alone, especially in light of the scarcity-attitude nature of (not) being recognised for your work in the field. Everybody is terrified that they will a) not be able to publish, b) be 'scooped' on their idea, or c) their months and years of often back-breaking labour will not be recognised or justified in light of their results - especially if those results are 'negative'. Which brings me onto...
2. The way academia works, only 'positive' results surface to the publication and sharing level of science.
This is limiting for two reasons:
- Nobody knows whether they are pursuing experiments that have, at some point, been carried out before. Therefore scientific work isn't cumulative: we are not building on the work of each other. Instead, we are sort of semi-cumulative: keeping an eye on what has been achieved/worked, and searching for the things not yet discovered.
- Whether something is a 'positive' or 'negative' result depends entirely on the eyes of the beholder: data and discoveries are not intrinsically positive or negative. It all depends on context. I guess this is a bit like the phrase 'one man's treasure is another man's trash'. Thousands of observations that, in the right context or setting, or interpreted by the right brain, may lead to amazing discoveries, just go to waste or lie dormant.
How can we create a culture of celebrating and rewarding all results, not just those that can, in the current paradigm and time stretch, be spun into a 'success' story?
3. By the time research finally reaches publication and public domain, the research going on in that lab may have already evolved, changed and leapt forward many times.
As Esvelt talks about in his article, this can also have implications for regulation -- in the current system, it will always be too slow. This is worrying. If we really are on an exponential trajectory of this field (genetic engineering, biotechnology, synthetic biology) evolving, then regulation will only continue to lag more and more behind.
It also has really scary implications for the consequences of new developments in science and technology. To lift the quote from the article from Oppenheimer, "When you see something that is technically sweet, you go ahead and do it, and you argue about what to do about it only after you have had your technological success."
We need to return the dynamic nature into science. We need to update, or create anew, the systems by which we share our data, ideas and build on each other's success and failures. We need special publications which aren't made for framing perfectly formed ideas and data, but for the brainstorming, hashing out of ideas and hacking problems that goes on in the entrepreneurial sphere. Maybe we need to run hackathons on scientific mysteries and problems. Or create publications that can be crowdsourced, crowd-written and editable.
4. Beyond lab groups, the public, and people from other disciplines and worlds of work, have very little way of finding out what a research group is currently working on, and what sort of opportunities for collaboration may exist.
How can we engage scientists from different silos to get involved in each other's work? In fact, how can we engage all people from outside of a scientific/technological field in that field?
I mean getting the designers and artists involved in synthetic biology. The philosophers in particle physics. The plant scientists in neurobiology. The lawyers in A.I. The historians in cancer biology. I know this happens already to some extent, but we kind of have media bias here: every time an artist does absolutely anything in a science subject, it's all over VICE and Vox and every other trendy magazine. The truth is, it happens not nearly enough and it's sort of bloody outdated that we all sit thinking about one specialised problem or field, and apply all of our thinking patterns and understanding in just one problem context.
I really like MIT Media Lab's approach in this area, and I think other institutions could learn from this. But let's not wait on entrenched academic institutions to change any time soon. Groups, collectives, departments, individuals can start de-siloing right away. Instead of referring to 'interdisciplinary', which refers to people from different disciplines working together, MIT likes to use the word 'anti-disciplinary' - in their words, they work "in spaces that simply do not fit into any existing academic discipline—a specific field of study with its own particular words, frameworks, and methods.”
With that in mind, MIT Media Lab has recently launched their Journal of Design and Science to explore and encourage antidisciplinary work coming out of its, and other institutions’, classrooms. The publication is very different to traditional academic journals -
There’s no anonymized peer-review process, and there’s no fee to access its contents. “We wondered what does an academic paper look like when it’s more about the conversation, and less about tombstones,” Ito says, referring to a quote from Stewart Brand that likens formal academic publishing to burying ideas like the dead. The journal is published on PubPub, a platform developed at MIT that is inclusive in ways that academia and academic publishing frequently aren’t; PubPub is an experiment in radical transparency, where almost every part of the journal is open and editable. Readers can annotate each paper, adding comments and context to what the author wrote. The editing history is visible to everyone, so authorship is no longer an opaque attribution. Hillis’ paper has executable code that can be lifted directly from the journal.
Ito describes the process as “peer-to-peer” review. The goal, he says, is for ideas presented in the journal to morph and evolve and become interconnected over time.
The way we create technology, do business, access education, and more is rapidly changing. We need science, and the way its done, to keep evolving, if we don't want it to get left behind. Millenials want to be part of the world that is built through collaboration, sharing, open-source and rapidly permuting to deliver solutions to our biggest problems. Let's not leave science behind.