Pros/Con of artificial intelligence - government organizations

Part
01
of two
Part
01

Artificial Intelligence - Benefits to Government: United States

Government departments that are immensely reaping the benefits of artificial intelligence (AI) technology include transport, healthcare, manufacturing, financial services, agriculture, weather forecasting, and national security and defense. The United States' government is creating a national climate where technologists and scientists can successfully create new AI inventions in the country. The Trump Administration is working on removing regulatory barriers to enable the development of new AI-based markets and the inclusion of AI by existing markets.

THE UNITED STATES' ARTIFICIAL INTELLIGENCE (AI) INITIATIVE

  • President Trump announced on February 11, 2019, the American AI Initiative to protect and promote national AI innovation and technology. Through this initiative, the government is partnering with the private industrial sector, the public, academia, and various international partners to implement AI technology in all its operations and organizations.
  • In this "whole-of-government strategy", Federal agencies will be required to independently or jointly pursue a multipronged approach in the advancement of AI in terms of promoting AI R&D investment, removing regulatory barriers, providing training and education opportunities to the country's workforce for AI, enhancing access to quality data and cyberinfrastructure, ensuring that the United States leads the world in the development of AI standards, and implementing a strategic plan to protect the country's advantage in AI.
  • Since 2015, the country's federal government has invested heavily in R&D of AI technologies and innovation. The growth of investment rose by 40% along with varied investments made in the field of intelligence and defense.
  • The sectors that are seen to benefit from the adoption and implementation of AI technologies are transport, healthcare, manufacturing, financial services, agriculture, weather forecasting, and national security and defense.

TRANSPORT

  • The government is implementing several autonomous systems in the transport sector including self-driving vehicles and unmanned aircraft systems that provide tremendous benefits to the country's society and economy. Through these systems, the delivery of household goods, mobility options for senior American citizens and individuals with disabilities, improvement in the safety of hazardous occupations, and the wider access to essential medical supplies will be transformed.
  • The Trump Administration, through NASA and the Department of Transportation, is partnering with State and local governments to implement AI both on roadways and in the United States' airspace.
  • According to Elaine L. Chao, the secretary of transportation, the safe deployment of AI technologies in terms of automated vehicles will ensure the decrease in traffic fatalities and improved mobility for the citizens.

HEALTHCARE

  • The Trump Administration is investing in AI technologies that will assist in the advancement of cutting-edge research in the medical field. The Department of Health and Human Services (HHS) is working to promote medical innovation via its DRIVe (Division of Research, Innovation, and Ventures).
  • The government is using AI to control diseases like Sepsis which claims 250,000 lives each year and is regarded as a national security threat. The HHS provides innovative solutions for diseases such as Sepsis by offering partnerships with private and public institutions. Additionally, HHS is working on Health Tech Sprint in an attempt to show how AI can be implemented into Federal data to develop products for healthcare applications.
  • The Centers for Medicare and Medicaid Services’ Center for Medicare and Medicaid Innovation has announced the launch of the AI Health Outcomes Challenge, in partnership with the Laura and John Arnold Foundation and the American Academy of Family Physicians. This collaboration is aimed at distributing up to $1.65 million to ensure further progress in AI for healthcare applications and to increase the rate of development of AI in real-world applications.

MANUFACTURING

  • The government, through the National Counterintelligence and Security Center, determined manufacturing areas such as space and exploration technology, integrated circuits, aircraft parts, advanced robotics, artificial intelligence, and chemical manufacturing as "areas under threat from foreign espionage".
  • President Trump has launched a National Strategic Plan on Advanced Manufacturing that primarily focuses on defending the country's economy, expanding employment in the manufacturing sector, and ensuring a strong defense industrial base with a resilient supply chain.
  • The National Institute of Standards and Technology (NIST) is working on transforming the current manufacturing sector into the smart manufacturing generation by developing autonomous, distributed sensing, and automated control systems. The primary goal is to improve the operational efficiency of manufacturing factories across the country.

FINANCIAL SERVICES

AGRICULTURE

  • The government is grappling with the biggest challenge in the agriculture sector — "feeding an additional 2 billion people by 2050". The government, through the Department of Agriculture, is investing in innovative technologies like AI to ensure food security. In addition, the department is working to technologically empower farmers, educators, scientists, and the public to benefit from AI.
  • The Department of Agriculture is also providing economic opportunity through innovation, helping the rural section of the country to thrive, and promoting profitable and efficient agricultural production.
  • Food and Agriculture Cyberinformatics and Tools is a data science initiative launched by NIFA to expand on AI-related programs that ensure numerous uses in workforce development through training and education, rural community support, ecosystem management, bioinformatics, sensor development, and agricultural production.
  • This initiative consists of robotic solutions that use AI technologies to help fruit harvesting, pesticide applications, weeding, and pollination; machine learning algorithms that help in identifying tree, animal, and plant species that contribute to ecosystem management and pest control; and adaptive watershed and groundwater models to maintain agricultural systems.
  • These investments contribute to a wide range of AI-relevant research which includes intelligent decision and computer vision support systems, machine learning, and big data.

WEATHER FORECASTING

  • The National Oceanic and Atmospheric Administration (NOAA) is utilizing AI technology to better understand and predict the country's dynamic environment. The NOAA is also using AI to improve nautical charts to ensure efficient maritime commerce, survey fish stocks so as to effectively manage the $208 billion per year commercial and recreational fishing industries, monitor the world's coasts and ocean conditions for national security applications, conserve endangered species, mapping, and exploring.
  • NOAA utilizes onboard machine learning techniques in satellites to assist in the protection of the environmental spacecraft during radiation occurrences.

NATIONAL SECURITY AND DEFENSE

  • National Security and Defense is one of the main sectors that have immensely benefited from innovative technologies like AI. President Trump announced a new National Security Strategy in 2017 that rallies the country behind innovation, technology, and research in emerging technologies such as AI.
  • In administration's commitment to AI is well pronounced in the National Defense Strategy, where the government outlines its intent to invest broadly in the military application of machine learning and AI along with other emerging technologies.
  • In June 2018, the DoD launched the Joint Artificial Intelligence Center (JAIC) "to serve as the focal point in the use of AI for key defense missions". The JAIC is tasked with delivering AI-enabled capabilities, synchronizing the various AI activities of DoD, and scaling the impact of AI.
  • Earlier this year (February 2019), the Department of Defense released its AI Strategy which focuses on harnessing artificial intelligence technologies to advance the security and prosperity of the United States.

Part
02
of two
Part
02

Artificial Intelligence - Downsides to the Government: United States

The downsides of artificial intelligence (AI) in US government organizations from the governments point of view include AI making government organizations vulnerable to security breaches, AI's self-reinforcing biases, government organization's incompetence in AI, the conflict of interest between government organizations and AI software manufacturers, and AI making it easier for individuals to lie.

GOVERNMENT ORGANIZATION'S VULNERABILITY TO ATTACK

  • The use of AI in US government organizations at the federal level faces security problems. Many government institutions in the US have faced security breaches at some point. This has been due to the lack of securing data in the organizations, allowing hackers to attack and punish the organizations. It has become increasingly unlikely for government organizations to protect algorithms from cyber attackers. It's also becoming difficult to identify if and when attackers have tampered with an algorithm, as algorithms are usually "black boxed."
  • The recent allegations of the hacking of the Democratic National Committee by Russia, which reportedly assisted Trump’s presidential bid, is a good example. The stealing of sensitive data for 21.5 million government employees and applicants after hackers accessed the U.S. Office of Personnel Management, is another example. Besides, coming up with laws that would aid in the transfer of judicial and government powers to artificial intelligence machines, gives rise to issues of software infiltrated with bugs.

SELF-REINFORCING BIAS

  • What AI does best is to analyze a lot of data in real time to identify what is going on and give conclusions on trends. The main problem with this is that AI has a self-reinforcing bias, as it assumes that all things happen the same and that there will be a repeat of what happened in the past. An example is the crime data produced by AI machines that single out minority and black neighborhoods with lower incomes, as they are more likely to be plagued with anti-social behavior and crime than prosperous white ones, for rigorous police patrols, thereby reinforcing the notion of biases by the police.
  • Al has been used in predictive policing at the local level in the US since the early 2010s, and has constantly faced accusations of being faulty and predisposed to deep-rooted racial biases. Self-reinforcing biases may be to blame in such cases as questions are asked on how transparent the AI algorithms are.

GOVERNMENT ORGANIZATION'S INCOMPETENCE IN AI

  • A downside of artificial intelligence in US government organizations is the lack of competence in digital matters. Many in government organizations lack experience in understanding the power and limits of AI technology. This means that their failure to grasp AI fundamentals prevents them from sufficiently regulating software companies that build AI software.
  • Some officials in US government organizations are incapable of understanding how AI functions in terms of things such as backdoors and their relationship with encryption, making it unlikely for them to sufficiently come up with regulations, such as algorithmic ones, for the industry.

CONFLICT OF INTEREST

  • There exists a conflict of interest between government organizations and AI software manufacturers in the US. These organizations need to regulate and police the AI industry using software systems that are not developed by them, but which are often developed by private tech multinational corporation. Questions arise about such partnerships and how transparent the algorithms sold to government organizations are.
  • Private tech multinational corporation have been accused of advocating for free and open data when it is coming from the government, but strongly opposed to the same concept when it comes to their data to ensure that it remains proprietary.

EASIER TO LIE IN AI ADAPTATIONS IN MILITARY OPERATIONS

  • The Department of Defense has adapted AI into military operations. Although AI is said to increase the sources in truth discernment in intelligence operations, its downside is that it has also been known to make it easier for people to lie persuasively. In addition, AI has been used to enhance forged video and audio, and brought into question what is authentic and can be trusted across many government organizations, including in the military.
  • The adoption of AI by the military poses a problem in that the current validation and verification process is not suited to AIs that learn, as it is developed for frozen software. In addition, corrupted data, possibly from enemies, could have grave consequences in military operations. It is also difficult to trust an AI system in the military that cannot be easily understood.

Research Strategy

    To provide the downsides of artificial intelligence in US government organizations from the governments point of view, our first strategy was to search for precompiled data from credible industry and academic databases, reports, and publications with relevant information to fulfill the request. This strategy led us to sources such as World Economic Forum, Research & Development, The RAND Corporation, Defense One, and McKinsey & Company, which provided data on the downsides of artificial intelligence in US government organizations from the governments point of view, which we have summarized above.
Sources
Sources

From Part 02
Quotes
  • "Virtually all major corporations, government institutions and agencies – including the US Department of Justice – have likely been breached at some point, largely because such organizations tend to lag far behind the hackers when it comes to securing data. It is, to put it mildly, unlikely that governments will be able to protect algorithms from attackers, and as algorithms tend to be "black boxed", it’s unclear whether we’ll be able to identify if and when an algorithm has even been tampered with. "
Quotes
  • "There will be pros and cons to the adaption of A.I. in military operations. The researchers said A.I. could increase the number of sources in intelligence operations to discern the truth, but also make it easier for subjects to lie persuasively. A.I.-enhanced forgery of audio and video media is rapidly improving in quality and decreasing in cost. A.I.-generated forgeries will challenge the basis of trust across many institutions."
Quotes
  • "However, there are also implications from AI adoption by the military. The military's current verification and validation process is meant for frozen software and is not suited to AIs that learn. Tainted data, possibly from adversaries, might have fatal consequences. It is also hard to trust a system that cannot be understood."
Quotes
  • "Too soon to mention the hack of the U.S. Office of Personnel Management, in which 21.5 million government employees and applicants had their sensitive data stolen, most probably by China."
Quotes
  • "When you think about the limitations, I would think of them in several ways. There are limitations that are purely technical. Questions like, can we actually explain what the algorithm is doing? Can we interpret why it’s making the choices and the outcomes and predictions that it’s making? Then you’ve also got a set of practical limitations. Questions like, is the data actually available? Is it labeled? We’ll get into that in a little bit."
  • "But I’d also add a third limitation. These are limitations that you might call limitations in use. These are what lead you to questions around, how transparent are the algorithms? Is there any bias in the data? Is there any bias in the way the data was collected?"