A race against time: The negative effects of omnipresent artificial intelligence

Twenty-odd years ago, tech enthusiasts were eager to point to Moore's Law as a way to predict the potential of a computer to achieve a human level of artificial intelligence (AI). They believed that they would see human-level AI by 2040 and super intelligent AI by 2060.

Today, society's AI dependency goes far beyond the smartphones, chatbots, and AI-assisted manufacturing and production of the early 21st century. And while machines haven't quite yet taken over, sectors like healthcare, retail, mining, and infrastructure use AI to conduct business and drive profits.

Though no one denies AI's valuable role in daily life, regulators are racing against time to mitigate the negative impacts AI's omnipresence is having. They have set up multilateral agreements to regulate global practice and are reaching consensus on issues that require regulation, including:

• How AI has led to greater income disparity and left a subset of the population denied the dignity that comes with work and a sense of contributing to the societal good.
• Which jobs, particularly those that require more nuanced skills such as patient care, should be done by humans. This includes measuring the ratio split between human and machine for certain jobs and roles.
• The misuse of broad surveillance and the weaponization of data that is harming underrepresented groups.
• Aiming to ensuring AI does not discriminate but rather promotes and supports diversity and inclusion.

There is also continuous debate amongst regulators and other stakeholders about whether AI fits within existing legal frameworks or if a new categorization with explicit features and implications must be developed. This is especially in light of AI assistants performing initial legal, financial, and ethics reviews that are only later checked by humans.

Not everyone agreed back in 2021 that AI was the technological savior the world had been waiting for.

There were many occasions where the use of AI had a negative impact on society.

• Some AI recruiting tools showed to be biased against women.
• False facial recognition matches led to arrests of innocent men and women.
• Faulty AI systems in self-driving cars have led to major traffic accidents and in some cases, death.
• GPT-3, an AI system has shown to be discriminatory towards Muslims.

Back in 2021, regulators were also struggling with global AI regulation. In April of that year, the EU drafted the first-ever legal framework for AI, encouraging the ethical development of AI technologies in the member nations. Still, differences of opinion and practice were already emerging, particularly when it came to the protection of fundamental human rights, such as the right to privacy.

Regulators in the EU set up and implemented the GDPR, or General Data Protection Regulation, the strictest data privacy and security law up until that time. It included hundreds of pages of requirements for organizations around the world that target or collect data related to EU citizens or residents. For instance, one article stated that a person had the right to not be subject to a decision based solely on automated processing, including profiling.

It's clear AI cannot be a no-law zone. An increase in AI regulation is a good first step towards offering a common minimum level of ethics and related penalties worldwide. It's encouraging that there are now standard levels of privacy, security, safety, fairness, accountability, and transparency that must be adhered to across multiple industries and sectors. But as new AI technologies are developed and released into society, further regulations should be crafted and imposed.

Moving forward, advancements in AI systems will likely make the technology even more omnipresent, affecting people's well-being and livelihoods in every imaginable way. Ethics advocates say they hope principles that focus on the public good will continue to play a critical role.