Ano’ng pwede natin gawin? AI is coming for your job
When the network giant GMA-7 presented to the public their newest sportscasters Marco and Maia, the Internet was all abuzz. Why? They look normal for sports reporters except for one thing—they are not humans.
GMA-7 calls the move “innovation in journalism” and reassured the public that the AI human models will not replace warm bodies in media. Many were not appeased and quick to voice out their fears—will these AI reporters replace their human counterparts?
4th Industrial revolution
Artificial intelligence, or AI, refers to the computer simulation of human intelligence processes. This involves programming machines to imitate human thought patterns by exposing them to diverse datasets. The goal is to replicate how the human brain works and apply it to the specific tasks the machine is designed for. The development of computer algorithms approximating human intelligence is made possible by the increasing computational power of machines.
The rise of AI is part of the ongoing fourth wave of industrial revolution that is taking over many aspects of our lives. We, as individuals and as part of society, are increasingly dependent on the internet due to easier access to information and services. AI products such as AI-generated images, videos, songs, and applications are becoming common. In the physical world, AI has allowed the development of self-driving cars, which used to exist only in sci-fi films.
Industrial applications of AI have also been on the rise. In the oil and energy industry, AI is used to predict the demand and needs for future use; in aviation, to plot more economical flight routes; in marketing, finance, e-commerce, and education, for predictive data analysis.
Among the most recent advancements in the field of AI are the large language models (LLM). These models are built on artificial neural networks—algorithms that mimic the structure of a human brain—and trained on a big data set of information. The result is a computer software that can carry out sophisticated conversations with human users, who may likely mistake it as a normal person.
Applications of these LLMs can be seen in OpenAI’s ChatGPT, Google’s Bard, and in Meta’s Llama, as well as in conversational AIs such as chatbots, virtual agents, virtual assistants, and even content-generation for SEO (search engine optimization) that are commonly used to assist in research, learning, or even for mundane tasks such as turning home appliances on and off (e.g. Siri, Alexa), or when you need someone to talk to in boredom (e.g. Simsimi).
Because of the language proficiency exhibited by these platforms, and the sheer volume of information that they can access and process, many think that these AI platforms can replace humans in some tasks. After all, AI integration has made work easier, and for some sectors, safer and cheaper.
And the tasks that can be delegated to AI has been growing. Even in the arts, AI has made its footprints. Image, music, and video generation can already be enhanced—if not completely and totally done—by AI. It can even write a full article! (This write-up, by the way, is not generated by AI.)
AI vs humans
Since new technology also impacts the way people work, it also brought disruptions to society and the economy. On the positive side, it improved productivity. On the negative side, it destroyed old ways of doing things where things are normally done more slowly, and more deliberately.
The advent of the internet and e-mail is why traditional postal service declined. In car factories, manufacturers opted for automation and replaced car painters with workers who know how to operate automated car body painting equipment; this raised production outputs, and cut the number of employees to be compensated.
The postal service example depicts direct displacement of economic activities, and in effect quite a number of workers. In auto manufacturing, the number for workers was reduced. New technology requires a new set of skills. Workers who were able to acquire these skills would remain on the assembly lines; those who failed to skill up were laid off.
However, despite the lay-offs and disruptions, it is important to note that automation has provided a much safer workplace for employees. It has reduced if not removed workers’ direct exposure to hazards, improving their health and welfare.
In the past, workers assigned to repetitive and routine-based tasks were deemed to be easily replaceable by machines. Many of these tasks can be observed in manufacturing (e.g. filling, canning, sealing) and agriculture (e.g. planting, harvesting, sorting), although some are also present in high-value added service activities such as BPO and finance (e.g. customer service and accounting tasks).
Cognitive and artistic work were thought hard to automate. But that was the view some years ago. Currently, even workers engaged in creative writing fear that they are as replaceable as machine parts. In fact, this is one of the grounds why the emergence of AI tools such as ChatGPT worries creative workers and has sent unions such as the Writers’ Guild of America (WGA) striking in May 2023.
Alienation of labor
Creative writing is not the only one threatened. The performing arts is also now disrupted. The means to produce images and performance in films through AI already exists.
In the US, film producers want to take advantage of this opportunity by “capturing” the likeness and performance of actors to train generative AI, which will then be used later in lieu of real actors.
Currently, CGI is being used to capture and replicate actors, where a body double is used in place of the original actor, and their faces to be placed with tracking dots to enable easier overlay of the original actor’s face. In the AI version, there would be no need for a body double.
Surely, the actors will get paid for their performance and whenever a production uses their likeness, right? Not necessarily. It all depends on who owns the captured performance and likeness of actors. Film producers assert their ownership of the captured likeness and performances of actors. As soon as images and motion have been transformed into data, it is now the AI taking over acting. Actors lose both jobs and earnings.
The use of AI is one of the issues that caused bargaining between film producers, represented by the Alliance of Motion Picture and Television Producers (AMPTP), and actors, represented by the Screen Actors Guild-American Federation of Television and Radio Artists (SAG-AFTRA), to break down.
SAG-AFTRA was forced to launch a strike in July 2023 to protect the interest of actors and secure work for them. On 9 November 2023 after more than 100 days, the union announced that it is calling off the strike after securing a better package for workers in the film industry. This includes "unprecedented provisions for consent and compensation that will protect members from the threat of AI".
Regardless of the arguments, one thing is certain: Whoever owns the technology also tends to capture the most returns from it—unless workers take collective action.
The future of work
The encroachment of AI in the world of work is just getting started, and technological advancement will continue. Although the impact on workers especially in developing countries such as the Philippines is not yet felt significantly, it will eventually.
In more advanced countries such as in the US, the possible impact of generative AI is already being considered as millions of jobs are aimed to be automated – which probably means that workers may be displaced.
Sam Altman, chief executive of OpenAI, has declared that “there will be an impact on jobs”. After all, in the event of mass automation, who shall be responsible to protect workers from possible displacement? While automation guarantees higher productivity, it will also create new types of jobs as it replaces existing ones.
Historically, automation is known to cause a decline in employment as displaced workers are not always readily equipped to take on the tasks demanded by the newly generated jobs. What can be done to address these concerns?
Alongside its goal to increase AI integration, Germany has put in place policies to aid workers from possible displacement due to AI. Some of the safeguards they have put in place are: Workers may be given paid leaves to attend educational courses; displacement tax are imposed on companies refusing to retrain their employees for AI-skill up; government offers financial incentives to companies that would enhance existing jobs to accommodate both AI and non-AI tasks.
In a recent discussion, Germany’s Research Minister Bettina Stark-Watzinger says that for 2024, additional AI skill centers and professorships shall be added to fortify their goal to boost AI integration in the country, as well as improve existing AI-infrastructure and worker training.
Which still begs the question: Who shall be responsible for the protection of workers?
As everyone is affected, everyone shares accountability. But most of the weight is on employers and government, as they hold this new means of production and regulate its use respectively.
This employer-government cooperation should identify pros and cons of such change; lay out long-term goals to achieve higher productivity, advancement, and economic improvement without sacrificing the welfare of the workers; and create and implement programs that would help them adapt to the fast-changing labor landscape.
What about us workers? We need to remind ourselves that we who provide either physical or intellectual or creative labor give things value. Our labor creates value. And the new machines should remain tools that make our labor less back-breaking.
References
Emeritus. 2023. Examples of Artificial Intelligence (AI) in 7 Industries.
New York Times. May 23, 2023. A.I.’s Threat to Jobs Prompts Question of Who Protects Workers.
An industrial policy for economic upgrading and achieving desirable social outcomes
Introduction
The fourth industrial revolution will bring radical changes—both positive and negative—in the economy and in society. Convenience and ease of carrying out tasks are among those that new technologies such as the internet of things (IOT), artificial intelligence (AI), blockchain, and automation, also considered disruptive technologies because of their impact on consumer and producer behavior. Consumer welfare increases because these new technologies facilitate consumption by reducing costs, and such is the case of online shopping where consumers realize huge savings in terms of time spent on shopping. Firms also benefit from better information and analytics brought by machine learning and artificial intelligence, and lower cost of production due to automation. However, the changes brought about by the fourth industrial revolution have serious drawbacks particularly for economic activities and jobs that will be destroyed. In the long run, huge returns accruing to the owners of technology and capital can reinforce existing income inequality. High levels of inequality can cause social division that can negatively impact the economy.
Although the impacts of technological change on jobs and income inequality have been explored in the literature, it is likely that these foreseen effects will occur unless insights on how to address the negative effects of technological changes are translated into actual policies. However, most policies that seek to address the negative implications of automation and disruptive technologies assume the inevitability of job destruction and focus on encouraging workers to acquire new skills. Skills upgrading of workers is important to keep them employed. However, a policy that mainly involves supply side approach that only responds to anticipated technological shocks will likely serve business interest more and put most of the burden on labor. For instance, although facilities for skills upgrading exist, it remains costly to undergo specialized trainings especially when it is left to market. Workers who cannot afford to undergo trainings will certainly be left behind by technological change. A better approach to the fourth industrial revolution involves the state intervening to influence firms to create jobs where technology complements labor rather than substitute for it, and encourage investment of firms on their workers through skills upgrading.
This policy note surveys existing government policies on the fourth industrial revolution and discusses insights on how the government can address employment challenges created by rapid technological change. This policy note also argues that social outcomes must be incorporated in industrial policies to ensure that 4IR does not lead to welfare losses for workers.
Government policies
The first explicit government policy direction in response to the fourth industrial revolution is documented in the Department of Trade and Industry’s ‘Inclusive Industries Innovation Strategy” (I3S). Based on the said document, the government’s strategy is oriented toward harnessing the possibilities of the 4IR, with the twin aims of (1) elevating priority industries’ position in the global value chain and (2) generating more and high-quality jobs for Filipinos via innovative technologies as key motivations.
According to de la Cruz (Forthcoming), skills development appears to be the most developed component of the government’s fourth industrial revolution response to date. The transition to 4IR is the major theme in the Technical Education and Skills Development Authority’s (TESDA) 2018-2022 National Technical Education and Skills Development Plan (NTESDP). Now that TESDA has been attached as an office directly under DTI, its skills development policies must now be closer to firms and industries. One of the NTESDP’s principal objectives is to “prepare the Philippine workforce for the challenges of the 4IR”, chiefly via the increased adoption of enterprise-based training and Skills Needs Anticipation mechanisms and stronger alignment of national TVET regulations to international standards (TESDA, 2018). Beyond the existing TESDA scholarships and job-specific apprenticeships, however, the issue of funding for inclusive worker re-/upskilling efforts remains in question. This uncertainty is exacerbated by inconsistent government support for such programs, as illustrated by the lack of government funds allocated to TESDA’s “Tulong Trabaho” initiative in 2020, which was supposed to provide free TVET training and related financial assistance to qualified beneficiaries, including those displaced by technological unemployment (CNN Philippines, 2019).
To realize its other goal of generating more formal and stable employment for Filipino workers, the i3S promotes the advancement of technology-driven growth and increased industrialization. Apart from the establishment of an innovation and entrepreneurship ecosystem, a crucial element of the i3S is thus the revitalization of the Philippine’s manufacturing base (e.g. electronics, automotive), given the sector’s greater capacity for job generation and stable working conditions compared to the services industry (Usui, 2012; Aldaba, 2013; Llanto and Ortiz, 2015). A major example of this support to the manufacturing industry is the Comprehensive Automotive Resurgence Strategy (CARS) program. First launched in 2015, CARS aims to position the Philippines as a regional automotive manufacturing hub by drawing new investments, stimulating domestic demand, adjusting relevant government regulations, and providing time-bound and performance-based fiscal incentives to domestic vehicle and parts manufacturers (Rosellon and Medalla, 2017). A more recent initiative is the Securing Manufacturing Revitalization and Transformation (SMART) program, which was announced by DTI in December 2019 (Manila Bulletin, 2019). The SMART program is set to provide financial support to priority projects in manufacturing such as electronic vehicles and rural industrial development, with the aim of encouraging the industry’s adoption of innovative technologies that will enhance its position in the global value chain and translate into job-creating growth.
Ensuring that social outcomes are achieved
Social outcomes must be integrated in industrial policies so the transition toward 4IR does not reduce welfare of workers. First, while awareness of the impacts of technological change is important in the transition toward 4IR, policies that provide support to firms and workers must be adequately funded. The government’s skills development plan is an important policy. However, unless such program is funded, training of workers will only have limited reach and produce suboptimal outcomes such as the mismatching of skills, or the outright shortage of workers that possess the appropriate skills for work.
Second, skills training must be matched by income support for workers. The government recently enacted unemployment insurance (UI) for workers in the private sector. UI, especially when adequately funded, can provide an income bridge especially for workers who are undergoing skills upgrading. Moreover, UI can also motivate unemployed workers to undergo skills trainings because they are assured of income despite not being able to work.
Lastly, employment creation must be part of industrial policies. Policies like CARS and SMART have the potential of generating employment by influencing firms’ decision toward adoption of labor-complementing technologies. However, the state must still ensure that firms that enjoy fiscal support indeed generate employment. To this end, an explicit employment target can be set for firms that will avail of fiscal support.
The implications of automation on inequality
Introduction
The fourth industrial revolution will bring radical changes to the economy, both positive and negative. The internet of things (IOT), artificial intelligence (AI), blockchain, and automation—these are just of the new technologies that are set to change the economy and people’s lives. Consumers will mainly experience positive changes due to these technological changes. Meanwhile, workers are among those mostly on the losing side especially when these technologies cause displacement and redundancies and force firms to reduce their workforce.
But not all technological changes are of the labor-displacing type. Acemoglu and Restrepo (2018) and Autor and Salomons (2018) identified four types of technological changes that can occur under the fourth industrial revolution and these can be characterized by their impacts on the labor market:
- Labor-displacing technological change depresses wages and decreases labor demand.
- Labor-augmenting technological adoption increases the productivity of labor, leading to increases in wages (but not employment).
- Intensive-margin advances deepen the productivity of capital or machines in areas where automation has already progressed, increasing both labor demand and wages.
- Creation of new tasks involves technological change that develop niches for activities where labor has a comparative advantage relative to capital, thereby boosting labor demand.
Current discourse on automation puts emphasis on labor-displacing technological changes because of their impact on employment, income, and inequality. Automation will render many tasks performed by labor redundant and this implies not only employment loss but also loss of income loss especially when existing jobs are not enough to absorb labor. Even if labor-displacing technological change occurs alongside creation of new tasks, movement of workers from old, obsolete jobs to new tasks is not guaranteed. First, it takes time for workers to acquire new skills and hence, to fit in new tasks especially when tasks require specialized skills. Second, the number of new tasks that can be created cannot also be guaranteed to match unutilized labor. What is more likely, however, with technological change is job polarization, i.e. the emergence of low skilled, low wage jobs that most likely belong to the informal sector.
The returns to technology will be captured by owners of capital. Huge returns to capital because of technological change will reinforce existing income inequality unless corrected (Korinek and Stiglitz, 2017). High levels of income inequality can be damaging to society and this has been explored widely in the literature. Without appropriate intervention to redistribute the returns to technology, the current trends in technology adoption under the fourth industrial revolution will certainly increase inequality in society.
Technological change, wages and labor share
The type of technological change occurring is difficult to determine. Moreover, existing data only points to its effects on wages, employment, and capital returns. However, these relationships can be used to infer what kind of technological change exists in an economy in general. This paper presents the results from an econometric exercise done to determine that relationship between technological change as represented by total factor productivity (TFP), and employment, wages and labor share to total output. The results presented here are obtained from fixed effects regression using firm-level data from the Annual Survey of Philippine Business and Industries (2013 – 2016). The econometric exercise yields three main results.
First, technological change has insignificant impact on employment. Technological changes in firms as measured by TFP does not affect the level of employment of firms, thus existing technology of firms do not cause labor displacement. However, this is not guaranteed to persist in the long run.
Second, technological change has positive impact on compensation. It is interesting to note that compensation expense of firms increase due to technology. Increase in compensation may be caused by improvements in productivity due to innovations in production.
Lastly, technological change reduces labor share of output. The output of firms is commonly measured as the sum of payments to labor (wages) and capital (rents). When labor share of output declines without wage cuts, then it must be the case that returns to capital are growing at a rate faster than that of wages. This result has long term implications on the level of inequality.
Addressing the negative impact of technology
Labor displacement and decline in labor share are often seen as natural outcomes of rapid technological change. Leaving the market on its own will not address inequality resulting from technological innovations. The state has an important role in determining the direction of technological change especially on how it affects society.
First, new tasks that emerge from new technology may require workers to acquire specialized skills. The state and intervene in this area by providing adequate support to workers as they undergo skills upgrading through access to trainings and income support for the unemployed.
Second, the state can encourage firms to adopt technologies that augment and complement with labor. Such technology will lead to increase in productivity of workers and at the same time, real wages.
Lastly, if inequality worsens because of huge returns to capital due to innovations, the state can correct such inequality through taxation. Technological changes occurred because society made investments in them, directly or indirectly. These investments may take the form of public goods that facilitate economic activities, or through an education system established by the state. Thus, society must also share in the returns to innovation.
COVID-19 Pandemic and Automation
Introduction
Not only did the COVID-19 pandemic challenged the otherwise great performance of the Philippine, economy, it also forces firms and micro entrepreneurs to adopt radical changes on how they conduct economic activities. Technological advancements particularly in machine learning, artificial intelligence, and information and communication technologies will facilitate rapid adoption of new ways of doing business especially for firms that can afford them. However, given the huge economic losses resulting from lockdowns and work stoppage due to the pandemic, firms may find it sound to invest in these new technologies if it means fewer economic losses while the pandemic is ongoing.
The COVID-19 pandemic
The government imposed lockdown in mid of March 2020 to stop the spread of the virus. What was supposed to be a 2-day lockdown was extended to two months. Before May 2020 ended, the government started relaxing strict lockdown measures to allow some businesses to resume because the enhanced community quarantine imposed on most parts of the country caused huge economic costs. Cruz et al (2020) estimates that about 20 billion pesos are lost every day of lockdown. In 2nd quarter 2020, GDP declined by 16.5 percent, the largest economic contraction since the Second World War. Coinciding with the economic crisis is decline in employment by almost eight million in April 2020, and involuntary hunger reaching a new peak in six years.
The government allowed economic activities to resume gradually with caution because workplaces can become COVID transmission hotspots immediately. Recovery was not uniform. While some sectors can immediately resume with minimal changes in conduct of business, other sectors needed major changes. Experts from the University of the Philippines came up with classification of sectors based on how important they are in economic recovery and the risk that they transmit COVID. Among those that are of high risk are workers in retail and recreation, transport and logistics, and public works and construction. All these sectors are important in economic recovery, however, because they involve close physical interaction, they also display high propensity to spread COVID.
Technological change under the pandemic
Suspension of economic activities due to COVID has caused businesses to resort to technology to adapt. Retail, which is one of the sectors severely hit by the pandemic, has shifted online as more consumers embrace online shopping. The emergence of online retail platforms and the growing popularity of online payment systems facilitated the shift of retail activities online. The growth of online shopping also encouraged growth in logistics as the increase in online purchases created jobs in the form of door-to-door delivery services.
Recreation and entertainment will also be different. Theater houses and music halls can easily become COVID super spreaders so it is unlikely for the film, music and creative industry to return to traditional means of service delivery. Some of producers in the sector have started looking into online showing of films and music performances. The technology that enables such form of distribution is already existing and will improve continuously as more creators and producers utilize them.
Radical technological change is also expected to occur in the construction industry. Construction activities are known to employ huge number of workers. Because of this, the sector is also considered at high risk of spreading COVID. Major firms in the construction sector are already eyeing more prefabrication of building parts through the design for manufacture and assembly (DFMA) system. Under DFMA, activities in the construction sector will be split into two processes that take place separately: manufacturing of building parts and structure, and assembly on site. With this system, labor costs of construction firms will sharply fall because a significant portion of construction processes will be taken care of by manufacturing and only the only tasks left for firms are those that involve assembly. The DFMA method is highly appropriate and will address the current challenges of the construction sector under the pandemic. Manufacturing of building parts and structure will greatly benefit from developments especially in 3D printing. Outsourcing tasks to manufacturing will reduce risks of COVID transmission in construction because physical contact will be reduced.
The employment challenge
Technological changes because of the pandemic are inevitable for adaptation and survival of firms and economic activities. However, they come at a cost especially to workers. The abrupt shift online will significantly reduce the number of workers employed especially in sectors such as retail. The adoption of new construction techniques will also leave many workers unemployed if they cannot be absorbed by new jobs. Although there will be job losses, there will also be new jobs created, although they will likely require new set of skills.
Skills upgrading is more important than ever for workers to be able to take on new jobs. The COVID-19 pandemic will accelerate technology adoption and automation, and unless workers are equipped with necessary skills, rapid technological change will result in inequities especially because the jobs most exposed to automation and obsolescence are those that are routine and require low skill sets usually carried out by workers in lower income groups. The state has to step in to address the adverse implications of technological change by providing means for workers to acquire skills through trainings, basic income for those that will be displaced by technology, and encourage firms to adopt technologies that complement rather than substitute for workers.
1 Based on April 2020 Labor Force Survey results and SWS September 2020 survey.