Rise of the machines: Adopting AI and its impact on skills in the research industry
1st March 2018 16:31
A blog by Liz Murphy, Senior Research Manager
Having binged on the new series of Black Mirror recently I thought I would continue the technological / futuristic theme a couple of weekends ago and got well acquainted with the robots currently residing at Manchester’s Museum of Science and Industry. The excellently curated exhibition traces robotic development from its earliest known origins at the beginning of the 18th Century (with commissions by the Catholic Church*) to the present day.
The exhibition leaves you with a real sense of optimism. After the insanity bender of 2016 and the ‘waking up the next day’ apprehension hangover that was 2017, starting 2018 with a spot of optimism felt quite nice.
Although it’s centred on machines, the exhibition does an excellent job at reminding its audience of how brilliant and complex humans are. It emphasises how difficult it is to accurately engineer and emulate our movements, facial expressions and understanding and appreciation of context. Video snippets of experts in the field of Robo-stuff (I’m pretty sure that’s the official term) do a really warming job of outlining the nuances of human physiology, how complicated yet natural our facial, hand and walking movements are, as well as the crazy amount of information we are able to digest and interpret in seconds. Subsequently, developers discuss how difficult it is to flawlessly recreate it all.
While they are able to build robots that can pick up pens, they find it difficult to engineer them to twist them in the miniscule and precise way that we do. Robots can be designed to lift and pull a pint, but it’s tricky to make them mirror the movements of drinking that pint. Walking has been a thing for them since the 1920’s, and these days, robots can recover when pushed over (aww, please don’t push over robots in the name of science) and backflip (sweet) but apparently engineers can’t quite replicate “stumbling naturally” yet.
So it seems that sadly, despite being in development since the 1700’s, we’re still not in a position where we can get robots drunk. I ask you. What has science been doing for 300 years?
Other than the physical aspects, humans are still winning at other things as well.
Identification for instance. Despite huge progress being made, current technology is yet to be able to accurately tell the difference between a large splodge of crème fraiche and a pile of powder, both being indiscriminately identified as a mound of some sort of white substance. A nightmare for those with robot-cleaners as apparently the two different substances require very different cleaning methods.
Interpretation? Absolute minefield. Adaptability? Big picture thinking, partnership working? Forget about it. Rumour has it that their knowledge of history is also not the best.
Throughout the exhibition, a number of issues were raised about the development and use of robots that were also relevant to the evermore prevalent discussions around the opportunities / threats of AI in general. Some of these issues were also touched on in Research Live’s ‘Skills and Talents Preview for 2018’, where industry leaders discussed the skills they felt to be key to the research world in this coming year. https://www.research-live.com/article/features/preview-2018-skills-and-talents/id/5032412
There was definite appreciation that AI should be embraced. But the weight of comments were given to the importance of “soft” skills. As well as the need for getting cosy and comfortable with technology, there was an emphasis on the importance of those skills that currently have not been mastered in machine-form. Empathy, flexibility, team-working and being able to identify long-term solutions to name a few.
Annie Pettit’s comment resonated with me the most:
“Our industry needs a new kind of unicorn, one that understands technology and one that is ethical and human. Technology is going to affect all of us and whether you’re an engineer, designer or questionnaire writer, we all need to speak the language of technology to ensure the work we do is done properly and efficiently. At the same time, our human qualities are becoming more important than ever. We’ve seen numerous cases were AI and automation vastly improved technical efficiencies. But, we discovered that those same systems also perpetuated racism and sexism. Humanism is essential to identify and prevent this from happening. We desperately need employees who are genuinely kind, caring, understanding and flexible.”
Pettit’s view triggered thoughts of the critique of IQ tests’ potential for cultural bias when measuring intelligence. As religion and Blade Runner will testify, when anyone creates anything, there is no escaping the fact that it is going to be created in their image. AI will only ever be heavily reflective of its creators and the values that they live by.
Technological advancement and change are inevitable. But we can still decide the direction and objectives of the change, the end goals and the values that drives those goals.
The end section of the exhibition showcases robots that are currently in operation across the globe. At the risk of seeming joyless, whereas Toyota’s trumpet playing robot and the “Robothespian” can be appreciated for their technological brilliance, I couldn’t ignore a feeling of “well...great…in a way, but in the grand scheme of things, he’s not Miles Davis”.
In contrast, Kaspar, designed to act as a social mediator to build confidence with autistic children, is hugely impressive…in the grand scheme of things. The former examples seemed to be engineered to replace humans, the latter towards advancing them. Understanding and advancing all humans. That’s not a bad end goal for AI.
Granted, the world has been divisive since the dawn of time. But currently it feels divisive at times when it doesn’t always need to be. Important issues have been raised over the last couple of years and they need discussing with care and humanity so they are not lumped into one homogenous, unidentifiable mound and tackled with no nuance. We are running the risk of retreating into defensive corners and battening down the hatches, rather than using our human, collaborative, currently non-replicable qualities to full advantage by trying to understand from each other’s perspective the underlying causes of the issues, what’s really going on and how best to tackle them.
We can still control AI’s trajectory, we just need to sort ourselves out first, decide which values should govern that trajectory and consider how the current climate will manifest itself in technological / AI development.
It was interesting to see that “optimism” was one of the skills to be valued by the industry this year. If you weren’t optimistically inclined, you might worry that that optimism is now considered to be a skill (hard or soft skill? Possibly yet to be categorised by the Department for Education…). The exhibition, however, provided lots of reasons to feel optimistic. The research industry doesn’t feel to be in grave danger just yet, humans are amazing, science is incredible and a Terminator-type scenario is probably not going to happen within our lifetime, as cyborgs can’t properly master pens yet.
*One of the first displays shows examples of the Catholic Church’s investment in automatons. It included a mechanical Jesus designed to roll his head to one side while his body sheds drops of wooden blood. A mechanical Mary reaches up to him…Just going to leave that with you.