Automating the World
Last year I met professor Jerry Kaplan, from Stanford, who published his new books on AIs (AI: a guide to the near future [1] and We don’t need people [2]) with my University. In a short, one-hour lecture, he managed to paint a very good overview of where AIs are now and in what directions they will go in the next 10-15 years. I draw inspiration from his dissertation for this little essay.
Machines, able to do a better job (in some fields) than most humans can do, have reduced costs, improved efficiency and made us wealthier as a society. Many things were human effort was taken as necessary have now been automated, which helped not only business (helping companies grow and succeed in an increasingly competitive global scenery) but, and it’s often overlooked or underestimated, also society as a whole (providing the general public with goods that were once only available to the most wealthy) and even factory workers (making the alienating toil of working on an assembly line a thing of the past, something to laugh at in Chaplin’s Modern Times).
Often thought as a way to replace human intelligence, usually with alarmistic intentions, AIs are better understood when we consider them as an extremely advanced form of automation. As it often happens with new technologies (from electricity to vaccines), the general public will react with a mixture of fear, incredulity and suspicion. Are AIs going to replace us? Will my job be made redundant?
In my opinion, the West should face the Digital Revolution with an open mind, indeed ready to face the new challenges AIs will pose to our way of life, but without resorting to self-destructive, and ultimately futile, neo-Luddite impulses. What western society needs to do now is to prepare for a world where, as it seems probable right now, jobs that don’t require human imagination or some kind of human interaction will get automated and replaced. According to a report by the McKinsey Global Institute, around 15% (400 million) of global workers may be displaced due to automation by 2030, even considering a midpoint, conservative estimate; a 3% (75 million) of the workforce could also need to change occupational category [3]. But is this a bad thing? Some jobs will become obsolete, but completely new professions will be created and other will be redefined: still according to the same report from McKinsey, an 8-9% of the 2030 workforce will be in new occupations.
Automation is not an exclusive of menial or repetitive jobs though. Recently, a Chinese AI named Biomind beat 15 medical doctors in a tumor diagnosing competition:
“When diagnosing brain tumors, BioMind was correct 87 percent of the time, compared to 66 percent by the medical professionals. The AI also only took 15 minutes to diagnose the 225 cases, while doctors took 30. In regards to predicting brain hematoma expansion, BioMind was victorious again, as it was correct in 83 percent of cases, with humans managing only 63 percent.” [4]
Will this mean that medical doctors will soon go out of business? I don’t believe so, at least not in the foreseeable future (and let’s not discommode visions of a post-scarcity society). What will probably happen (as many experts agree on), when this and similar technologies hit hospitals all over the world, is that AIs will be able to screen incoming patients, reliably identifying different pathologies and then separate those who need immediate assistance from those who do not, assign the perfect doctor for every patient, assist the medical équipe during surgery itself and so on. This isn’t sci-fi: automation in the operating room is becoming increasingly common [5], and patients who undergo robotic surgery are less reliant on opioids and painkillers when recovering from their operations [6].
This was just an example to show how AIs and autonomous technologies are not and should not be relegated to supply line optimization, stock trading or assembly lines development, but they’re getting closer and closer to the general population and to everyday fruition, impacting all kinds of different fields and environment. What we are experiencing and we are going to face is more akin to the last industrial revolution than a simple productivity boost.
Will some people lose their jobs? It’s possible, yes. But the same thing happened when the petrol engine was introduced, and we don’t need to ride horses anymore. What governments and companies need to do now is to prepare to re-skill an entire generation of workers, and to make so that the boost in productivity due to automation does not benefit only a selected few, without indulging often misguided and auto-celebratory socialist policies: I don’t believe implementing “taxes on robots” will really solve this issue, and neither will a too generous UBI (at least not in the next few decades); education considering a rapidly changing job market should be the primary concern of today policy-makers.
After all, innovation does not stop for regulators.
Sources:
[1]: (in Italian) www.luissuniversitypress.it/pubblicazioni/intelligenza-artificiale
[2]: (in Italian) www.luissuniversitypress.it/pubblicazioni/le-persone-non-servono
[4]: www.xinhuanet.com/english/2018-06/30/c_137292451.htm
[5]: www.intuitivesurgical.com
Users who have LIKED this post:
5 comments on “Automating the World”
Comments are closed.
I like the part the discussion about the mindset of people from the west would be like facing this updating technology. The social background of this cutting-edge technology is also a big part to consider and check. I also believe that the “dark factory” is gradually dominating the industry which may wipe out some jobs for like technicians due to machines’ high efficiency, preciseness, low energy cost, etc. A lot of problems would be posed after this unstoppable trend, and we should take all accounts of people’s reactions into consideration for the future generations to develop faster and more smoothly.
We’re definitely going to see some huge changes to the industry: the challenge will be not to allow tech disruption to turn into social disruption.
I hope regulatory policies will work more on a preventive/educational aspect of the problem more than a taxation/prohibition one
I really enjoy your articles using past examples to illustrate how AI will impact us in many aspects. However, I just have a different opinion as opposed to regulatory policies will work on AI. As far as I concern, AI is not regulatable; it holds supremacy on almost every single aspect of humans, especially the intelligence. If so, it would be impossible to regulate something who is a lot smarter than the policymakers.
Ai is not stoppable, just let it be. It’s just a new species that deems to win over us, according to Theory of Evolution by Darvin.
I have high hopes for AIs in the future as well, but I think we’re still a bit far away from the Singularity. I was more talking about AIs as advanced automation machine more than sentient beings
I agree that AI is here to enhance us, and not to replace us. Machine learning only provides a foundation for building understanding by learning from data models, and yet it still cannot generate complex explanations nor make inferences from higher-level knowledge. I guess we’ll see! 🙂
Users who have LIKED this comment: