Artificial Intelligence (AI) is not magic. It is not a new technology either. Underlying component technologies are available and also affordable. In isolation, they have little capabilities. But once they are fine-tuned and fused with science to create a human-like intelligence, AI becomes a powerful tool. It can show a magical performance. AI has a transformative power disrupting almost all major sectors. Intelligent policies are in demand to turn this transformative potential into a blessing for all. Hence, AI policy response demands a long-term perspective.
AI is the fusion of already available technologies. Some of them are sensors, data analysis, and interpretation techniques. A few of them are statistical inference, computational pattern recognition techniques like a neural network, image processing, programming and software engineering, and decision science. But once they are fused with appropriate perfection, human-like intelligence could be added to machines. The purpose is not to make those machines to be rivals of humans. Instead, to empower machines to perform the job better, causing less wastage and harm. With this added capability, the industrial economy moves to a new height, called Industry 4.0 or the fourth industrial revolution.
Significant target areas for AI innovations
Not only factories, starting from healthcare to national security to agriculture, but all significant areas of our society are also subject to AI benefits and suffering. It’s predicted that AI is perhaps the only technology in recent memory that will defy all the hypes. To our surprise, it will likely turn out to have been underhyped once its impacts fully unfold. In a September 2018 Washington Post piece, Nicolas Berggruen and Nathan Gardels shared a strong message. According to him, AI has the unfolding potential to become the most powerful resource. It will likely determine the fate of nations in the times ahead. As observed by John Villasenor of Brookings, “It’s an astute observation that will likely prove true.”
Over the decade, it is an assessment that AI will be strongly linked with economic prosperity. AI will also become linked with geopolitics to a level that is difficult to comprehend today fully. Some areas where AI will have transformational changes are 1. Food production, processing, and supply, 2. Industry, 3. Finance, 4. National security, 5. Healthcare, 6. Criminal Justice, 7. Transportation, 8. Smart cities, 9. Education, and 10. Pollution and Climate change. There are numerous examples, such as semiconductor processing or food inspection, where AI is already making an impact on the world and augmenting human capabilities in significant ways. To leverage them, we need a long-term AI policy response.
AI innovations are progressing
USA, China, Russia, and the UK, among other nations, have come up with significant plans and investments. Among others, the financial sector has been showing a strong response to leverage AI. For example, investments in financial AI in the United States went up three times between 2013 and 2014 to a total of $12.2 billion. According to observers in the financial sector, software decisions about loans. AI software can take into account a variety of finely parsed data about a borrower; for sue, it is far more than just a credit score and a background check.
In healthcare, AI tools are being developed to analyze medical images. Despite imperfection, the autonomous detection of suspected regions by AI algorithms in the Computer Tomography (CT) images shows real potential. AI algorithms are showing potential in criminal justice, such as ranking people using age, criminal activity, victimization, drug arrest records, and gang affiliation.
There has been a race in developing self-driving vehicles to bring revolution in the transportation sector. Research of the Brookings Institution has found that over $80 billion was invested in autonomous vehicle technology between August 2014 and June 2017. Smart city initiatives are taking place all across the world. The development of AI solutions is underway to exploit the potential of precision agriculture to produce 15 to 25 percent more food from the same land using far fewer inputs than before. Similarly, factories are ready to benefit from AI innovations to produce higher quality products causing less pollution.
Every industrial activity starting from fabric production to food processing is a candidate to improve performance by 20 to 25 percent. Summarily, AI capability is ripe for harvesting to transform the education sector. There are already concerns about unmanned weapon systems and cyber-attacks in national security.
No barrier to access essential AI technologies
Is there a technology barrier to exploit the underlying potentials in making our diverse activities smarter? Is the technology only accessible to a specific group of countries or firms? Interestingly, the answer is no. All the underlying technology components, starting from sensors to processors, are commercially available; and they are also very low cost. For example, a powerful image sensor or processor to process image data costs less than $100. Underlying scientific knowledge is also available as the public good. From an overall perspective, AI is at the early stage of the technology lifecycle.
There appears to be no barrier for anybody to access basic technology and scientific knowledge components. But once those technologies are refined with science and fused to address specific issues, a powerful solution emerges. For example, self-driving cars’ major ten technologies are commercially available. But the refinement and fusion of them can lead to far safer roads saving lives and properties. Besides, it will likely be causing disruptive waves, transforming automobile and transportation industries.
AI innovations are elusive—AI policy response must take into consideration
AI appears to be highly elusive in terms of technology potential. The integration of available hardware and software components quickly leads to concept demonstration. Such crude forms often create an elusive impression that AI solution substituting human intelligence is just about the corner. For example, developing a mobile robot with essential obstacle detection and avoidance capabilities seems to be undergraduate project work. Over the last two decades, many such undergraduate projects have been completed. But the journey of fine-turning this capability in building an autonomous vehicle having level 5 autonomy is a daunting task. As reported, upon investing $80 billion, we still have to go a long way to turn the potential into reality. Many of the AI innovation potentials are like this, creating an illusion of reaching AI singularity.
AI policy response should focus on managing a long journey
It’s a long journey to turn the potential into useful innovation. Financing this long journey is beyond the reach of start-ups and private initiatives. There is a need for long-term research finance to support the refinement and fusion. Some of those initiatives even may demand research and development efforts running over decades. In the absence of public funding, such potentials will mostly remain untapped.
Moreover, the market for early-stage solutions should also be created. Public procurement policy may also need to be adjusted to create the market. Besides, AI policy response on the transformation of research into innovation should get due priority.
AI is a labor substituting technology
Developing countries should also rethink their technology policy. The policy of importing technology and adding labor and raw material is losing its ground in the age of AI. There is no denying that AI has an intense labor substituting capability. There is a prediction that a 50 percent chance AI will outperform all human tasks in 45 years and automate all human jobs in 120 years. Developing countries should focus on developing local innovation capacity to create employment for innovating AI solutions.
Moreover, as component technologies are readily available at a very low cost, it’s an opportunity for developing countries to enter into an innovation economy by leveraging AI. Moreover, AI also allows accumulating monopolistic market power—demanding smart policy responses. Hence, AI policy response should maintain a delicate balance.
AI policy response for bearing fruit and having balanced transformation
As AI is a transformative technology, there will be significant job loss in one area while creating new job opportunities in other areas. To manage this change, public policy has a substantial role to play. Developing the capability to leverage AI is a multi-decade endeavor—a far greater time scale than the term lengths of elected officials in democratic countries. Hence, implementing AI-focused policy strategies that might take several years or more to bear fruit may lower incentives to democratic governments. For the foreseeable future, AI will likely have a significant transformative impact on jobs and productive activities. It can create new opportunities if policymakers choose to harness them. But that requires intelligent, long-term policies running over decades. For minimizing stress on society, AI policy response should focus on managed transition, lasting over decades.