Artificial intelligence and machine learning technologies have permeated mainstream discourse in recent years. AI applications like ChatGPT and chatbots have taken the public by storm with their novel functions and potential applications.

However, many experts have also raised concerns about the potential risks of AI. Some public and private institutions currently proposed regulations for AI use and creation. These regulations would ideally help control AI in a way that protects human interests.

Read below to learn about the current state of AI regulation and the challenges that come with it.

What is AI regulation?

AI regulation involves the creation of government and institutional policies to regulate artificial intelligence. 

Due to the increasing adoption of AI in business and other industries, governments and regulatory bodies are taking notice. In the United States, for example, the Chamber of Commerce began calling for AI regulation in March this year.

Industries that have adopted AI systems in some capacity include healthcare, science and technology, insurance, finance, logistics, retail, and marketing.

Since the adoption of AI technologies is still relatively new, regulation is still up in the air. However, some governments and institutions have proposed guidelines on the responsible use of AI. 

Reasons for AI Regulation

Why does AI need regulation? AI applications have become helpful in various fields. However, there is still a lot of conversation about the risks they entail. The mitigation of these risks can help make AI safer and more productive to use.

While fear-mongering is rarely productive, the public and various stakeholders must take these concerns seriously. Some risks surrounding current AI technologies might make appropriate regulation necessary.

Privacy issues

Even before the widespread adoption of AI, many people have had growing concerns about data privacy and cybersecurity. 

In 2022, there were 1,802 cases of data compromises in the United States alone, affecting 422 million individuals. Data breaches can be dangerous, as they can include private information that leaves victims vulnerable to attacks like identity theft.

Some AI models use sensitive and private data to train their algorithms. Some companies use AI to collect audience and customer data to improve their marketing campaigns. 

While this feature can be helpful from a business perspective, some consumers are wary of what information these companies and technologies might have on them. Hackers and malicious actors can also use artificial intelligence to hasten their processes.

Human interactivity

Artificial intelligence changes over time, especially when humans interact with it. This aspect offers a lot of advantages and potential for innovation. However, there may be cases where AI can cause harm to people.

For example, in 2018, a self-driving car killed a pedestrian in an accident. The backup driver was charged with negligent homicide. 

AI is only a tool, and at this point, it can be risky to leave possibly life-altering decisions to it alone. AI-operated machinery and vehicles still require human intervention and consistent monitoring to ensure it doesn’t cause harm.

AI bias

Some people believe that AI is unbiased, as it is a product of a scientific process. However, this assumption could not be further from the truth. 

Computer scientists train AI using information currently available to and created by humans. Humans choose the datasets fed to AI and may intentionally or unintentionally cause the AI to internalize this bias.

In 2020, Michigan State Police wrongfully arrested Robert Williams, a Black man, for a robbery. Williams was in police custody for nearly 30 hours despite being innocent of the crime because local law enforcement’s face recognition software allegedly could not tell Black people apart. 

Without diverse teams and extensive reviews, AI could cause harm, misinform, or unfairly judge groups of people in various situations. 

Legal responsibility

When an AI tool makes a mistake, who becomes responsible? It can be challenging to identify the responsible parties when a supposedly independent AI causes harm or significant damage. 

In the case of the self-driving car discussed earlier, the human backup driver was deemed responsible. However, other AI applications can have completely different circumstances surrounding them. 

Similar conversations are occurring surrounding AI art. Some AI art generators have used copyrighted content in their output, which poses a significant concern regarding copyright infringement. In January 2023, Getty Images sued Stability AI for unlawfully copying and processing their copyrighted images.

Challenges of AI Regulation

One of the reasons AI regulation isn’t as widespread is due to the challenges surrounding the process. Regulating AI is not as easy as it sounds, especially as this technology constantly evolves.

Many industry groups, lawmakers, and governments are taking steps toward crafting regulatory documents for AI. However, these are some of the most significant challenges they must face throughout the process.

AI is a powerful tool that can change and grow as it sees fit. While many have used AI to improve business and organizational processes, poor data and biases can sometimes harm various groups and individuals.

Establishing ethical guidelines can help ensure no individual or group uses AI in a way that hurts others. However, groups with political or commercial interests can make this effort a tad challenging. 

Some ways to improve the responsible use of AI technologies include the following:

• Prioritizing data privacy 
• Ensuring the transparency and explainability of AI training data and processes

Technical challenges

AI is still a relatively new technology. For this reason, it can be challenging for policymakers to create comprehensive guidelines for their regulation.

Different AI models can change over time as they learn and develop. It can be challenging to standardize regulations on something that has the potential to adjust independently.

Balancing innovation and regulation

AI’s inherent capabilities to make judgments and adjust accordingly provides fertile ground for innovation in various fields. Creating hard legal regulations without appropriately weighing the risks and benefits of AI may hinder the technology’s progress.

Where AI Regulation Currently Stands

Despite the challenges of AI regulation, some policymakers and various organizations have crafted proposals and documents geared toward the responsible use of AI. 

These regulations and laws cover risk management surrounding AI technologies while supporting innovation. 

Canada

In June 2022, the Canadian Parliament drafted a regulatory framework for AI called the Artificial Intelligence and Data Act (AIDA). AIDA aims to promote the responsible development of AI, participate in the global AI landscape, and protect Canadians from possible risks.

It uses a risk-based approach that aligns with proposals from other regional and international bodies and existing Canadian legal frameworks. Its three main principles include the following:

• Ensuring that high-impact AI systems adhere to existing consumer protection and human rights law
• Giving power to the Minister of Innovation, Science, and Industry to administer and enforce AIDA to help ensure the act evolves according to changes in technology
• Prohibiting malicious and reckless uses of AI through new rules and criminal law provisions

China

China doesn’t yet have laws and regulations specific to AI technology. However, they have laws regulating private companies’ use of online data for consumer marketing. 

Companies must conspicuously inform users whenever they use algorithms for content recommendations. Users can then opt out of being targeted with recommended content by algorithms.

In September 2022, Shanghai became the first province to pass a law on AI development. This provincial law aims to promote the high-quality development of AI technology. 

European Union

In 2020, the European Commission published a white paper entitled “On Artificial Intelligence – A European Approach to Excellence and Trust.” The paper called for a prompt initiative to support responsible AI development and ensure the technology serves the welfare of the European people.

In April 2021, the EU introduced the Artificial Intelligence Act. It also employs a risk-based approach with three main risk categories:

• Applications with unacceptable risk 
• High-risk applications
• Applications not banned or considered high-risk

Another priority of the EU is ensuring that current and future AI policies adhere to General Data Protection Regulation (GDPR) principles. This regulation covers concerns about data privacy and protection.

United States

There are currently no federal laws specific to AI in the United States. However, public and private institutions and local governments are working towards creating regulations to mitigate AI risks.

The Biden administration and the National Institute of Standards and Technology (NIST) have published guidelines for the safe use of AI. The White House has also released the AI Bill of Rights to address the misuse of AI and offer recommendations for its responsible use and implementation.

Using AI Regulation To Maximize Benefits for Humankind

As with many new technologies, it can entail a lot of work for governments and other entities to catch up in terms of laws and regulations. Creating advanced yet trustworthy AI technologies is a challenging yet necessary balancing act. 

Proper research and impact assessments can help in many governments’ and organizations’ decision-making processes regarding AI governance. A comprehensive AI strategy on regulation can help manage AI’s capabilities, improve human processes, and minimize harm.

Content Guide Outline

AI Regulation: How Do They Work?

Opening paragraphs

• Briefly introduce the article.
• Discuss how artificial intelligence has hit the mainstream and influenced various industries.
• Being a new technology, various bodies are calling for its regulation.
• Transition to the rest of the article.

Keywords to incorporate: ai regulation, artificial intelligence, ai use

[H2] What is AI regulation?

• Define AI regulation and discuss the conversations around it.
• AI regulation involves the creation of government and institutional policies to regulate artificial intelligence.
• The United States Chamber of Commerce has called for AI regulation in March 2023.
• Discuss the industries where AI is mostly used.

References:

• reuters.com/technology/us-chamber-commerce-calls-ai-regulation-2023-03-09/
• https://www.alston.com/en/insights/publications/2022/12/ai-regulation-in-the-us#:~:text=Industries%20Impacted%20by%20Increasing%20AI%20Dependencies%C2%A0

Keywords to incorporate: ai regulation, artificial intelligence, ai technologies, ai systems, initiatives, ai applications, chatbots, chatgpt, healthcare, science and technology

Reasons for AI Regulation

• Briefly introduce the section.
• Various leaders and groups want to regulate AI to mitigate its risks.
• Discuss the common reasons for AI regulation in the H3s below.

Privacy issues

Some AI models use sensitive and private data to train their algorithms.

Human interactivity

• AI has the potential to cause harm the more people interact with it.
• Discuss examples of cases where unchecked AI has led to harm to a person.

AI bias

AI can be biased through the data used to train them.

Legal responsibility

Who is responsible when an AI commits a mistake or causes harm?

References:

• https://www.tableau.com/data-insights/ai/risks#risks
• https://www.techtarget.com/searchenterpriseai/feature/5-AI-risks-businesses-must-confront-and-how-to-address-them

Keywords to incorporate: ai regulation, ai risks, ai development, high-risk, datasets, decision-making, cybersecurity, machine learning, mitigation, real-time, law enforcement, training data

Challenges of AI Regulation

• Write a brief introduction to the subtopic.
• Regulating AI is not as easy as it sounds, especially as it is constantly evolving.
• Discuss each challenge to AI regulation in the H3s below.

[H3] Ethical and moral considerations

Developing a set of ethics for developing AI can help minimize harm, though it can be difficult to monitor in practice.

Technical challenges

• Different AI models can change over time as they learn and develop.
• It can be difficult to standardize regulations on something that can change independently.

Balancing innovation and regulation

Some parties fear that regulation can hinder further innovation in AI.

References:

• https://www.ibm.com/topics/ai-ethics
• https://fortune.com/2023/04/03/how-to-regulate-ai-challenges-three-experts/#:~:text=Human%20foibles%20and%20a%20moving%20target

Keywords to incorporate: regulating ai, stakeholders, ai governance, policymakers, providers, ai regulation

Where AI Regulation Currently Stands

• Explore the current discussions on AI regulation.
• Discuss current AI regulations worldwide in the H3s below.

Canada

The Canadian Parliament drafted a regulatory framework for AI in June 2022.

[H3] China

China currently regulates how private companies use AI for marketing purposes.

European Union

The EU introduced the Artificial Intelligence Act in April 2021.

United States

The current administration and the National Institute of Standards and Technology (NIST) have published guidelines for the safe use of AI.

Keywords to incorporate: lawmakers, white paper, European commission, gdpr, ai policy, risk-based, biden, bill of rights, new rules, regimes, general data protection regulation, legal framework, impact assessment, human rights

Using AI Regulation To Maximize Benefits for Humankind

• Wrap up the article with a brief summary of relevant points.
• AI regulation can help manage AI’s capabilities, improve human processes, and minimize harm.

Keywords to incorporate: trustworthy ai, ai regulation, ai strategy, risk management