This article is written by Naufal Virindra, an Indonesian civil servant working on the Indonesian Civil Service Commission (KASN RI). This article represents personal ideas and not the institution.

  • The problem: There is a lack of talent and regulatory issues to control AI.
  • Why it matters: AI development is rapid and unavoidable.** **
  • The solution: Practical actions based on the OECD Principles on Artificial Intelligence.

This article proposes several recommendations for containing AI threats based on the OECD principles on Artificial Intelligence. This principle encourages AI to be innovative and trustworthy and to respect human rights and democratic values.

Jack Ma v.s. Elon Musk: Different views on AI

The debate about Artificial Intelligence (AI) future capabilities is intriguing, especially among technology giant founders like Jack Ma and Elon Musk. Jack Ma is the co-founder of Alibaba Group, one of the largest e-commerce businesses, and Elon Musk is the co-founder of several tech companies such as SpaceX, Tesla and The Boring Company. From the product point of view, we can see that they share the same enthusiasm for developing technology businesses. However, they have recently argued about the future of AI and its impacts on human life at the World AI Conference in Shanghai, China. Elon argues that AI can make jobs useless since humans think slower than computers. He argues that humans ponder in kilobits per second, while computers can take it to a terabit level. Jack Ma shares this opinion that computers may be clever, but human beings are much smarter. He thinks computers only have chips and that man has the heart as the source of wisdom. These differing opinions might show their standings regarding AI as a threat or an opportunity. The contradiction between these two tech tycoons is somewhat similar to China and the United States' current intense situation where both countries are superpowers in the economy, technology development and military firepower but are so different in political and legal systems, role of government and culture.

Artificial Intelligence: A computer made to achieve human-like ability

AI is a technology that enables computers and machines to generate human intelligence and problem-solving capabilities. It works by processing large amounts of labelled training data, learning the data for correlations and patterns and using these patterns to make predictions about future projections. Here are some breakdowns about AI cognitive skills and applications.

💬 Let the author know what you think about the differences between AI and huma intelligence by leaving a comment below

AI Cognitive Skills and Applications
AI cognitive skillsAI applications
  1. Learning: dedication to gathering data and formulating rules to convert it into actionable insights. These rules, known as algorithms, give computing devices step-by-step instructions for performing specific tasks.
  2. Reasoning: selecting the appropriate algorithm to achieve a desired outcome.
  3. Self-correction: continuously refine algorithms to ensure they yield the most accurate results possible.
  4. Creativity: employs neural networks, rule-based systems, statistical methods and other AI techniques to create new images, text, music and ideas.
  1. Speech recognition: uses natural language processing to convert human conversation into written format.
  2. Customer service: virtual agents or chatbots doing service by answering frequently asked questions.
  3. Computer vision: computer's ability to learn information from digital images, video, or other visual inputs. App examples include photo tagging in social media, radiology imaging in healthcare and autonomous cars.
  4. Supply chain: helping supply chains to react to machine-generated, augmented intelligence while providing instant visibility and transparency.
  5. Weather forecasting: weather broadcasters rely on complex algorithms run on supercomputers to make accurate forecasts.
  6. Anomaly detection: explore large amounts of data and discover atypical data points within a dataset.

These days, the use of generative AI is becoming more widespread. It is a deep-learning model that can gather all raw data, collecting, filtering and analysing everything available on the search engine. Deep learning is a sub-discipline of AI with a neural network composed of an input layer, three or more (but usually hundreds) hidden layers where data processing and extraction occur and an output layout. These multiple layers enable unsupervised learning, which automates the extraction of features from large, unlabeled and unstructured data sets. One of the famous examples of generative AI is ChatGPT (currently the GPT 4 version). People can easily access the website and have a casual chat with an AI about everything. ChatGPT will answer each prompt, and in some cases, it will share the data source. However, although ChatGPT seems reliable and easy to use, be careful with the answers given for every prompt. AI can also make mistakes. Therefore, Artificial Intelligence users are strongly encouraged to have a critical mind and verify through multiple resources to get the ideal and valid data.

Artificial intelligence examples: United Kingdom and Singapore cases

Using AI in government might sound fancy, but some countries have already utilised it. In the United Kingdom, the National Health Service has made a national Covid-19 Chest Imaging Database. It consists of a shared library of chest X-rays, CT scans and MRI images to help test and develop AI to treat Covid-19 and many other health conditions. It also means AI will use all data in the database to learn and train itself to bring the UK government's desired outcome in the healthcare context. In Singapore, there was Ask Jamie, a virtual assistant that supports citizens and businesses with government services across approximately seventy government agencies with AI-powered chat and voice. However, this chatbot type was recently replaced by VICA (Virtual Intelligent Chat Assistant). It is a government platform that takes benefits from Artificial Intelligence, Machine Learning and Natural Language Processing (NLP) to provide a public-facing chatbot. With these technologies, VICA enables one to learn and understand a conversation and continuously adapt to improve user interactions.

The threats

Although AI might offer many benefits, improve the quality of government services and provide skilful tools, as explained in the above examples, there is also a potential threat if the government fails to overcome several barriers to effectively implementing artificial intelligence. The first potential barrier is the lack of skilled talent. It is a bold move for the government to introduce AI in public service, but it will become a ludicrous decision if the government cannot provide a workforce that understands AI and how to train it. Government AI can potentially be underdeveloped or become a burden on regular work, which means it does not give any value to the government. The second barrier is the unclear regulations to make sure AI is ethical, secure, transparent and human-centric. It will be disastrous, as predicted by Elon Musk if there are no clear regulations to control the development of AI. The performance of AI can exit social norms (e.g. using bad words and gestures in communication), develop and train itself with irresponsible data (sharing hoaxes and false data that can affect government performance) and grow super rapidly until humans cannot control it (it becomes super intelligent and takes over humans primary work and life).

Solutions based on OECD Principles on Artificial Intelligence

Despite its benefits and threats, artificial intelligence development is certain and unavoidable. Therefore, it is important to consider living in harmony with AI, especially in the government.

OECD Principles and Recommendations
OECD Principles on Artificial IntelligenceRecommendations
  • AI should benefit people and the planet by driving inclusive growth, sustainable development and wellbeing.
  • AI systems should be designed in a way that respects the rule of law, human rights, democratic values and diversity. They should include appropriate safeguards – for example, enabling human intervention where necessary – to ensure a fair and just society.
  • There should be transparency and responsible disclosure around AI systems to ensure people understand AI-based outcomes and can challenge them.
  • AI systems must function in a robust, secure and safe way throughout their life cycles and potential risks should be continually assessed and managed.
  • Organisations and individuals developing, deploying or operating AI systems should be held accountable for their proper functioning in line with the above principles.
  • Continuously give, train and assess AI data to ensure optimum performance by providing a valid outcome. AI training is necessary to update the latest valid information, learn from past mistakes and improve AI-based outcome quality.
  • AI could make mistakes. Therefore, human intervention to ensure information validity in every source is crucial to prevent AI from giving misinformation or irresponsible outcomes.
  • AI training for government employees to understand how to operate, train, assess and fix the AI programme.
  • Regulation flexibility and upholding the core values of the OECD principle. Flexibility is required to develop AI that will always change, rebrand and renew to improve its version rapidly. However, some work and studies encourage originality. There is also a technology that can analyse whether some work was created by AI (for example, ZeroGPT). It can harm users' reputations and credibility if there's any plagiarism found or similar outcomes generated by AI. Therefore, AI users must obey the terms and conditions at work, in their studies and in their daily lives before using AI.

Done reading? Make sure to share your own thoughts on the OECD Principles on Artificial Intelligence by leaving a comment below ⬇️

(Image credit: Unsplash)