This topic emerged through a screening of emerging issues in science, technology and society in the context of the FET CSA OBSERVE.
The big picture
New forms of machine-human-symbiosis emerge on all levels and across types of activities. Aspects range from automation in all spheres of human activities to augmentation of intimate functions within the human body.
The OBSERVE Screening brought up the following elements:
More and more decisions and activities are done by algorithms and machines (hiring, investment, robo-journalism, poetry, recognition and reproduction of facial expressions, identification). Deep learning further extends the range and depth of human machine interactions.
Modelling the human
Technical abilities increasingly allow artificial imitation of the human body or behaviour. These imitations are used for testing human related artefacts. Developments include: Artificial/3D printed skin, chips that mimic human organs, real looking animated human flesh or artificial nerve cells.
Automation is fast progressing in ever more domains of human activity (medical re-search; sewing, software managing processes, trash collection, drone ships, automated hiring; automatic ordering/replenishment; robo journalism; science). This comes with a number of social, economic and ecological consequences. Humanity needs to understand and actively shape this development e.g. by controlling the power of algorithms and learning to deal with unexpected events. One of the key challenges will be to build a workable future for all.
The technological singularity is a hypothetical event in which artificial general intelli-gence would be capable of recursive self-improvement and thereby ultimately surpass human control or understanding. This expectation has long been voiced in Science Fiction but is also expected and discussed by several researchers. More and more such as physicist Stephen Hawking warn that the development of full artificial intelligence could spell the end of the human race. FET proposals point out that groundwork in fields like ontologies, knowledge representation, computational linguistics and cognitive systems is still necessary.
Various technologies, techniques and drugs are emerging, that can enhance bodily functions and in particular cognitive performance (neuroenhancement). There are mi-crochip implants, smart contact lenses which can monitor blood values or advanced prosthetic technologies (bionic eyes, extreme knees…). Severable applications serve to support blind and deaf people to navigate. Biohackers modify themselves with RFID-Chips or other technologies. Science fiction writers envision even more extreme forms by envisaging such as chips that can set emotional mind states or interact with body cells e.g. fighting cancer cells. Also artworks like “Anke” (Hans Op de Beeck): investi-gate the theme of the enhanced superhuman and the quest for immortality. This trend (in particular cognitive enhancement) could change the way we work, we think or how we communicate.
Brain interfaces and implants
There is a lot of current R&D effort at the intersection of neuroscience, biotechnology, and computer science directed at developing interfaces to the brain. The topic was one of the most prevalent in the webmining. Examples for research activities on interfaces are memory chips to improve the memory performance, brain controlled bionic devices and a brain reading technology that helps violinists to play again after a heavy accident. For implants new developments in 2015 included soft devices to deliver drugs in the brain and syringe-injectable electronics.
A new generation of robots with cognitive planning and reasoning capabilities is being developed. They can handle uncertainty, act in messy unpredictable situations and carry out creative tasks like participating in a cartoon contest. Furthermore some re-searchers focus on a kind of artificial consciousness and self awareness - a concept that is highly contested by some philosophers.
Virtual Personal Assistant Bots
Deep machine learning gives rise to a spectrum of smart machine implementations — including robots, autonomous vehicles, virtual personal assistants (VPAs) and smart advisors — that act in an autonomous (or at least semiautonomous) manner. Apps as virtual assistant such as Microsofts Cortana, Google Now or Apples Siri will get better in recognizing situations and derive individual user needs.
Fully autonomous production organism
If industry 4.0 becomes bigger and more connected, this could mean the end for industrial production as we know it. In science fiction it is envisaged that all production is accomplished by an artificial organism that works fully autonomously.
Rise of the drones
Science fiction novels envision a world where drones of all shapes and sizes will take over a vast diversity of functions such as monitoring, scanning, surveying, transport and spying. Smarter and smaller drones with more functions are one of the most popular areas on Kickstarter.
Robots will become more human-like as their vocabulary comes closer to that of real humans
Currently, computational intelligence is restricted to lexical descriptions found in dic-tionaries. Several FET projects however argue that with current results from brain research from clinical studies and neuro-imaging, a complete inventory of words, their emotional valence and perceptual properties will become available. However, it takes a combined effort of neuroscientists, cognitive scientists, lexicographers and computer scientists to harness this new source.
Cognitive overburden through perpetual evaluation
Much of today's software casts the user in a role of perpetual evaluation from which it’s relatively easy to escape (with one or two clicks) but impossible to reject all together. This may be a taste of what seems likely to become a potentially overwhelming routine feature of daily life in the near future. Each individual act of evaluation is trivial but the aggregative cognitive burden likely isn’t.
The human brain in the digital society
Researchers reflect on the question how the digitalisation of society affects the human brain. As an example some speculate that autistic behaviour patterns may be becoming more prevalent in the digital society
Insights from cognition research and biology may enable better Ambient Intelligence (AmI) systems
Context awareness and activity recognition are key components of the vision of Ambi-ent Intelligence (AmI). A common problem is that activity recognition can only be achieved in narrowly defined sensor configurations. Bringing together insights from biology, machine learning and control theory will allow new systems which take advan-tage of sensing modalities that happen to be available, rather than forcing the user to deploy specific, application dependent sensor systems.
Implants that store and transfer data
Prosthetics or implants which store, process and transfer data in many ways emerge in science fiction. These implants could be interconnected and also linked with a global net. In this scenario complete surveillance could be permanent.
Ideas about optical implants emerge in science-fiction. Nerve cell-prosthetics which translate a digital optical input in biological data or implants which replaces organs are thinkable. Furthermore there could be laser-implants transmitting additive information direct to the visual nerve via a laser-stimulus.
Automated indoor farming
A company in Japan is building an indoor lettuce farm that will be completely tended by robots and computers. The company expects the factory to open in 2017, and the fully automated farming process could make the lettuce cheaper and better for the environ-ment. For now, the Wall Street Journal reports that the company is still working on a machine that can plant the seeds, and their process still requires human eyes to determine whether a seedling has sprouted.
The work needed
Emergence of new forms of human machine symbiosis clearly carries substantial potential for benefiting humanity. At the same time several sources point to severe threats. Collaborative and inter-disciplinary research is required to reap the benefits, prevent failures and chaos and to counteract risks. Issues like trust and responsibility in hybrid human-machine ensembles need to be explored on all levels.