The European Union is poised to enact the European AI Act, a first of its kind for AI systems in the West, and a guide the use of artificial intelligence.
It's a powerful precedent and focuses on ethics, transparency, and rights protection, according to technology innovator and author Hassan Taher. The EU's actions, he notes, allow for AI to continue to progress as a technology while limiting the associated risks.
"I believe that AI technology has the potential to bring about significant positive change in the world, but many people are hesitant to embrace it fully," Taher notes. "While some may disagree, I believe that with responsible use, AI can actually make the world a better place for everyone."
The EU's Artificial Intelligence Act Explained
In June 2023, the European Union approved draft legislation that's considered to be the world's first comprehensive AI law. It would create a rule book for the adoption and use of this technology throughout the EU's 27 member states. The draft version of this law will now be negotiated with the member states and the Council of the European Union before it becomes law.
The act takes a risk-based approach to AI. It proposes a ban on what it deems the most hazardous AI practices, including the use of real-time facial recognition in public spaces.
It would also ban AI systems identified as intrusive or discriminatory by the EU Parliament. Such technologies include social scoring systems and models that use "subliminal or purposefully manipulative techniques."
It would enact stricter guidelines for generative AI models, which use neural networks to find patterns in existing data. Generative AI models, such as the popular ChatGPT platform, use this data to produce new, original content.
Under the EU legislation, generative AI platforms would need to disclose that content is machine-generated and build safeguards into their models that prevent the creation of illegal content.
Risk is central to the EU's artificial intelligence legislation. It seeks to categorize AI systems with regard to their potential threat to users. It categorizes AI into four levels of risk: unacceptable risk, high risk, limited risk and minimal or no risk
Under the proposed law, any AI systems deemed to have an "unacceptable" risk level will be strictly prohibited with few exceptions. Among those considered unacceptable under the AI bill are those that:
Manipulate people or use cognitive behavioral techniques. An example would be voice-activated toys that encourage children to engage in dangerous activities.
Social scoring that classifies people based on personal characteristics, socioeconomic status, or behavior.
Biometric identification systems used either remotely or in real time. Facial recognition tools are one possible example.
The draft bill calls for an exception that permits the use of remote biometric ID systems for prosecuting serious crimes when identification occurs following a significant delay; however, even such cases would need court approval.
Hassan Taher Applauds Basis of Trustworthiness in European Law
Taher is all in on the EU's actions. He praises the law for its commitment to safeguarding fundamental human rights while allowing innovation to continue. Taher is the founder of Taher AI Solutions, which provides guidance and strategic advice to companies that seek to implement AI across industries.
"Europe has emerged as a frontrunner in promoting the ethical development and use of AI technologies," he wrote in a blog post. "With a deep commitment to safeguarding fundamental rights, the European Union has taken significant steps toward establishing a robust regulatory framework. This framework aims to address the potential risks associated with AI while fostering innovation and long-term sustainability."
After studying computer science at the University of Texas at Dallas, Hassan Taher began writing about AI. He's the author of The Rise of Intelligent Machines, AI and Ethics: Navigating the Moral Maze, and The Future of Work in an AI-Powered World.
"By prioritizing transparency, accountability, and the protection of fundamental rights, the act aims to strike a delicate balance that ensures AI benefits society while minimizing potential harm," he recently wrote.
A Three-Tiered Approach to AI Regulation
Hassan Taher sees three important, fundamental components in the proposed legislation: trustworthy AI, an innovation-friendly landscape, and unified rules across the European Union.
By focusing on the need for AI systems built with ethical principles of fairness, robustness, and transparency, the proposed law has the right approach, Taher argued.
Taher noted that the law mandates the use of high-quality data sets, urges human oversight of AI models, and promotes explainability. "By mandating the use of high-quality data sets, encouraging human oversight, and promoting explainability, the EU aims to build trust among citizens, consumers, and businesses, ultimately fostering the adoption of AI technologies across various sectors."
The emphasis on protecting privacy and fundamental rights is at the core of the EU bill, Taher said. Europe aims to reduce the risks associated with AI by promoting privacy-by-design principles and mandating risk assessments.
In addition to listing unacceptable technologies, the identifies high-risk AI systems. These include toys, cars, biometric identification, employment, law enforcement, and critical infrastructure management. Other items on the EU's high-risk list are AI systems that could potentially influence voters during political campaigns and the recommendation systems used in social media applications.
All of these high-risk platforms would need to be registered in an EU database.
Even platforms deemed limited risk, such as ChatGPT and Google Bard, would need to adopt stronger protections under the new EU legislation.
These safeguards allow users to make informed decisions about if and how they engage with AI systems. Users would need to be notified if they are interacting with AI and given the chance to stop or continue using the AI system.
The law would also require stricter transparency procedures.
Taher also believes that the EU bill's approach allows for innovation and responsible use of AI while requiring developers to follow ethical guidelines. With clear standards in place, the EU legislation establishes a legal framework while encouraging innovation and investment in AI technologies.
The AI Act is a proactive approach to rule alignment that promotes uniformity and prevents fragmented guidance across different EU states. It also levels the playing field for developers and users alike.
Regardless of what changes it may undergo, the legislation will simplify how businesses in the EU comply with laws and encourage collaboration across borders.
Taher noted that there is an inevitability to AI. It is here to stay and will be disrupting businesses of all kinds.
Naturally, the industry's response to the proposed legislation has been mixed. For companies that are headquartered elsewhere but do business or have users in the EU, the guidance could complicate development and operations.
Nevertheless, guidance is necessary as AI technology has grown quickly and dramatically. In mere months, it's begun to transform and change how work is done across many industries.
The EU legislation allows companies to unlock the full potential of AI within an ethical, transparent framework.
It also reinforces a unified European ecosystem for AI, Taher noted. It establishes a cohesive approach to regulation and AI governance while encouraging transparency and accountability. The legislation is likely to become a model that other jurisdictions use when developing their own AI laws.
"As we move toward a future powered by AI, Europe's commitment to responsible regulation paves the way for a sustainable and inclusive digital transformation," Taher noted.
