By Marek Havrda, PhD, Olga Afanasjeva, and Will Millership
GoodAI
Intro
The emergence and global spread of COVID-19 has already changed the world in ways that were unimaginable at the start of the year. However, in many areas, it seems to have sped up changes that were on the horizon anyway in particular as online service, artificial intelligence, and technology, in general, become more advanced and more accessible. In this blog, we explore some of the trends of a future with autonomous and intelligent systems and how they link to the disruption caused by COVID-19.
1. State aid and restructuring the workforce
One of the major responses to the outbreak of COVID-19 has been a need to keep firms and households afloat during this difficult period. We have already seen special subsidies towards wages such as the Kurzarbeit in Germany which roughly translates as “short-time work,” or the furloughed scheme in the UK where the government will pay 80% of wages for workers on temporary leave. These schemes are only temporary but with the rise in automation, it is possible that mass layoffs of workers may become more prevalent (we have discussed this in more detail here). The pandemic has not only led to the faster deployment of various robotic solutions such as Blue Ocean Robotics whose robots perform various tasks such as disinfecting hospitals, but also to general automation of production of goods to reduce dependence on human workers. The pandemic will speed up not only the automation of manual tasks but also cognitive tasks such as those related to accounting, marketing, and HR. In order to deal with a new reality in the labor market due to a further rise of automation, governments will need to look at more permanent schemes such as universal basic income or conditional income linked e.g. to retraining. At the same time, governments will need new ways to fill the public purse. This could speed up tax reforms including a higher tax on the beneficiaries of automation and lower value-added tax on human labor-intensive products and services in the future.
Indeed, a progressive government could use this time window when costs of labor are subsidized due to COVID-19 for an aggressive transition towards more efficient automation-based production of consumer goods. This would allow us to decrease reliance on global supply chains and therefore become more self-sufficient on a national level (see also the next section). However, any transition would have to be accompanied by mass retraining schemes to avoid serious social decay. The retraining programs could, for example, focus on moving workers from the hospitality and travel sectors towards health, social care, and education sectors. AI could be used to improve the effectiveness and efficiency of these schemes, for example, Sana Labs is reskilling flight attendants to help the health services. The workforce restructuring could be sped up as the government may gain control over in-debted corporations (example: Alitalia) Governments could become visionary temporary owners of bailed-out corporations who would invest additional public resources into the reskilling of employees of bailed-out companies.
Additional pressures on the labor market may come from the increased concentration on various markets as large companies (often cash-laden tech companies) which will probably purchase various smaller cash-starved businesses and start-ups. This will bring additional efficiency gains leading to more layoffs and less choice. Competition authorities at the national level will need to be vigilant as this situation may bring out new challenges for competition policy.
2. Self-sufficiency, deglobalization, and the return of productivity
With the disruption to the global supply chain, countries may see the need to become more self-sufficient, particularly when it comes to strategic goods such as food, energy, health technology, and daily products. According to Rosemary Gibson of The Hastings Center at Yale University and the author of China Rx, the world is highly dependent on China for many generic drugs. This has been the case during the COVID-19 pandemic as Esther Choo, an associate professor of emergency medicine at Oregon Health & Science University, has pointed out, hospitals in the USA are already running out of vital medicines which are “needed to induce and maintain sedation while on a ventilator.” According to the Council for Foreign Relations, “about 80 percent of active pharmaceutical ingredients (API) are thought to come from China and India.”
The push to gain more independence will not be only relevant to “strategic” products and therefore supported by national governments. Also, private companies themselves will seek to reduce its exposure to risks inherent to multistep, multicountry complex supply chains. This need could also contribute to the faster implementation of automation through investment in more high-end robots, in order to keep production costs to a minimum.
In sum, we will see much more automation co-driven by self-sufficiency in more and more areas. As suggested in this article the need for self-sufficiency could see things like vertical agriculture (the practice of growing crops in vertically stacked layers to save space) become possible at community or country level, and as technology becomes better the vertical farms will become more efficient. The Financial Times recently highlighted the example of Keihanna, in Japan, which they called “one of the world’s most sophisticated examples of a vertical farm,” which can produce around 30,000 lettuces a day at full capacity. Agriculture like this, which is highly precise and efficient could help nations become self-sufficient in the future. Overall we may lose some efficiency derived from international trade today, but see more efficiency from automation and also more stability and control over supply chains.
If the above means more resilience remains an open question. However, we may expect a positive impact on overall productivity. The United States and other countries are experiencing low gains in productivity particularly in the last decade. One of the reasons may be that the delocalized production of goods has reduced the physical access of innovators to production reducing the opportunities to innovate. Much more automated re-localized production may prompt a return to productivity and growth. A further boost in productivity may stem from large-scale public expenditure (in R&D and beyond) in climate-friendly technologies including clean energy production allowing to operate data processing, cloud services, and AI systems economically and ecologically.
3. The move to remote
As COVID-19 threat became more serious almost all institutions had to adapt to remote life, and quickly. It has not been just offices and schools but even online gym sessions, private tutoring, and therapy to name a few which have boomed together with share prices of related companies such as Zoom. During quarantine, people became more accustomed to using online solutions through forced adoption. Companies that can provide online services are thriving without expensive marketing campaigns leaving behind off-line competitors.
Even galleries have gone online such as Art Basel Hong Kong fair which went virtual or a virtual spin around the Sistine Chapel. In the future, we could be experiencing even more things remotely. For example, using hi-fi virtual reality to buy a house, furniture, or a car. These changes will enable deeper personalization, allowing for more flexibility as people will be able to do a host of different activities from anywhere in the world at any time. Creators will also be entering directly inside our homes through platforms such as substack.com or patreon.com which enable direct connection with audiences.
Education, which has been resistant to change in many countries, has also had to be shifted instantly onto the online world. Many parents are complaining about the time demanding nature of homeschooling, however, this could be an opportunity for a step-change in education but also for a whole new staffroom of personal AI-enabled human-like tutors. At least for the time being it looks like AI will be soon good at creating personalized curricula. However, we know that well-performed one-to-one instruction is part of the most effective approach to education. This will speed up the development of human-like AI online teachers who will master and provide good old mentoring and motivation to “their” students. AI will also produce immersive learning experiences where students will be solving “real-world” problems in a virtual online environment which will much closer resemble the physical world so the learning experience will get much closer to learning-by-doing.
Also retail will speed up its transformation, not only in terms of the growing sharing of e-commerce but also in terms of blending e-commerce with brick-and-mortar stores, for example allowing the contactless checkout with mobile payment. Most of the sales of tickets for entertainment, travel etc. will be online, while cash payments will further decline. All of this will generate a lot of new data that can fuel various AI systems to optimize and improve products and services.
4. Collaborative science?
In the global fight against COVID-19, we witness a move towards open-sourcing solutions and open data which could benefit people. For example, there are open-sourced solutions for 3D printers allowing people to produce DIY personal protective gear. Even large corporations and organizations are opening up. As pointed out in this article Chinese company Baidu has opened LinearFold, its RNA prediction algorithm, so that researchers across the world can use it and it claims “to reduce the prediction time of a virus’s RNA secondary structure from 55 minutes to just 27 seconds.” The International Telecommunication Union (ITU) and the World Health Organization are working together with the help of UNICEF to deliver important health-related messages to people across mobile networks. Also preprints of academic articles posted at GitHub or preprint archives like bioRxiv (life sciences), medRxiv (medical), PsyArxiv (behavioral sciences), SocArXiv (social sciences), ArXiv (o.a. physics, mathematics, computer science) and Open Science Framework (OSF) or Zenodo increase speed and access to newly generated knowledge. This kind of collaboration and openness is promising especially when it is related to global health issues. However, we should not simply expect openness when it comes to strategic advantage, both commercial and military. For this kind of openness to materialize, it will require immense effort, visionary leaders, and international negotiations.
5. New security balance
Social mass monitoring has been used by China to help tackle COVID-19. They have been using data from drones, CCTV cameras, and even directly monitoring individuals’ movements to track the spread of the disease. Other countries like Israel, Singapore, and South Korea “are also using a combination of location data, video camera footage, and credit card information,” in order to effectively monitor the virus. Although these approaches may be effective there has been serious concern about the types of data collected and used, specifically biometric data, and how this could be used in the future.
With the increasing power of technology, it is highly probable that an increasing number of decision-making processes will be automated by the use of algorithms. For instance, Yuval Noah Harari gives a hypothetical example warning of how biometric data about your emotional reaction to a dictator’s speech could be instantly applied to adjust your credit scoring. Your data could be used in almost real-time affecting various aspects of your life, for instance, your ability to buy travel tickets, perhaps even your personalized pricing of certain services or goods to make them more or less accessible. In the future, because these processes will be automatically governed by algorithms, decisions about individuals could be made across populations.
There are emerging solutions to deal effectively with privacy protection while allowing for the societal beneficial use of personal data such as MIT Private Kit and there are many academics trying to support the creation of various decentralized solutions such as Personal Data Stores. The danger of the erosion of privacy will likely prompt a boom in the market for “white hat hacking” against surveillance. People will be developing apps and services that protect the data collected on personal devices leaking straight to governments or big corporations.
But in the closer future, we may witness the demand from current political elites in a number of countries for control of their population. As Richard N. Haass expects “many countries will have difficulty recovering from the crisis,” potentially leading to many more failed states. Their leaders will try to deploy various technologies such as tracing apps used to fight the virus to control their citizens and at least temporarily maintain some order.
We need to ensure that human-centered technology is not only an abstract concept. For example, to ensure that people are in better control of their own data and can decide where to send it and at what time. During a pandemic, you may choose to send your Fitbit data (which you send in peaceful times to your personal training program app) to a non-profit or even a government who is monitoring the spread of the disease, provided there are measures in place to secure such data. A good example of a human-centered approach to contact tracing has been put forward by the Harvard academics but we also need sustained monitoring of potentially privacy-invading measures used by governments and companies as outlined by Privacy International or the Computational Privacy Group. The crisis is also an opportunity for current giants such as Google and Facebook to strike a new balance with their users allowing them to control their data as well as with other businesses allowing them to strive using the new “digital utilities” into which the online platforms have matured.
Conclusions
Overall, the COVID-19 pandemic has already had a profound impact on society and is making us rethink the future. It seems to have sped up economic and societal transformations that would have otherwise taken years or even decades. The crisis is both a challenge and an opportunity.
Speaking from our own humble experience of AI R&D group, we noticed that with remote work becoming the new norm, we and people around us, are much more open towards remote collaborations, countering the popular belief that they’re intrinsically short(er) of human connection. We trust that this culture shift will fuel collaborative science even more, which is vital for overcoming the crisis, developing new scientific breakthroughs, and emerging stronger.
A lot of the outcomes will depend on how long the crisis eventually lasts. In some areas, we will get “back to normal,” while in many others we’ll need to adjust to a new normal. Some areas may significantly change as outlined above. In parallel with deglobalization, we might see global power shifts with the emergence of new economic and political forerunners, and the decline of established leaders. A lot will depend on how individual governments are responding to the pandemic, and economic and also social recovery.
