Artificial intelligence (AI) continues to maintain its prevalence in business, with the latest analyst figures projecting the economic impact of AI to have reached between $2.6 trillion and $4.4 trillion annually. 

However, advances in the development and deployment of AI technologies continue to raise significant ethical concerns such as bias, privacy invasion and disinformation. These concerns are amplified by the commercialization and unprecedented adoption of generative AI technologies, prompting questions about how organizations can regulate accountability and transparency. 

There are those who argue that regulating AI “could easily prove counterproductive, stifling innovation and slowing progress in this rapidly-developing field.”  However, the prevailing consensus is that AI regulation is not only necessary to balance innovation and harm but is also in the strategic interests of tech companies to engender trust and create sustainable competitive advantages.   

Let’s explore ways in which AI development organizations can benefit from AI regulation and adherence to AI risk management frameworks: 

The EU Artificial Intelligence Act (AIA) and Sandboxes  

Ratified by the European Union (EU), this law is a comprehensive regulatory framework that ensures the ethical development and deployment of AI technologies. One of the key provisions of the EU Artificial Intelligence Act is the promotion of AI sandboxes, which are controlled environments that allow for the testing and experimentation of AI systems while ensuring compliance with regulatory standards. 

AI sandboxes provide a platform for iterative testing and feedback, allowing developers to identify and address potential ethical and compliance issues early in the development process before they are fully deployed.  

Article 57(5) of the EU Artificial Intelligence Act specifically provides for “a controlled environment that fosters innovation and facilitates the development, training, testing and validation of innovative AI systems.” It further states, “such sandboxes may include testing in real world conditions supervised therein.”  

AI sandboxes often involve various stakeholders, including regulators, developers, and end-users, which enhances transparency and builds trust among all parties involved in the AI development process. 

Accountability for Data Scientists 

Responsible data science is critical for establishing and maintaining public trust in AI. This approach encompasses ethical practices, transparency, accountability, and robust data protection measures. 

By adhering to ethical guidelines, data scientists can ensure that their work respects individual rights and societal values. This involves avoiding biases, ensuring fairness, and making decisions that prioritize the well-being of individuals and communities. Clear communication about how data is collected, processed, and used is essential. 

When organizations are transparent about their methodologies and decision-making processes, they demystify data science for the public, reducing fear and suspicion. Establishing clear accountability mechanisms ensures that data scientists and organizations are responsible for their actions. This includes being able to explain and justify decisions made by algorithms and taking corrective actions when necessary. 

Implementing strong data protection measures (such as encryption and secure storage) safeguards personal information against misuse and breaches, reassuring the public that their data is handled with care and respect. These principles of responsible data science are incorporated into the provisions of the EU Artificial Intelligence Act (Chapter III).  They drive responsible innovation by creating a regulatory environment that rewards ethical practices and penalizes unethical behavior. 

Voluntary Codes of Conduct 

While the EU Artificial Intelligence Act regulates high risk AI systems, it also encourages AI providers to institute voluntary codes of conduct. 

By adhering to self-regulated standards, organizations demonstrate their commitment to ethical principles, such as transparency, fairness, and respect for consumer rights. This proactive approach fosters public confidence, as stakeholders see that companies are dedicated to maintaining high ethical standards even without mandatory regulations.  

AI developers recognize the value and importance of voluntary codes of conduct, as evidenced by the Biden Administration having secured the commitments of leading AI developers to develop rigorous self-regulated standards in delivering trustworthy AI, stating: “These commitments, which the companies have chosen to undertake immediately underscore three principles that must be fundamental to the future of AI—safety, security, and trust—and mark a critical step toward developing responsible AI.” 

Commitment from developers 

AI developers also stand to benefit from adopting emerging AI risk management frameworks — such as the NIST RMF and ISO/IEC JTC 1/SC 42 — to facilitate the implementation of AI governance and processes for the entire life cycle of AI, through the design, development and commercialization phases to understand, manage, and reduce risks associated with AI systems. 

None more important is the implementation of AI risk management associated with generative AI systems. In recognition of the societal threats of generative AI, NIST published a compendium “AI Risk Management Framework Generative Artificial Intelligence Profile” that focuses on mitigating risks amplified by the capabilities of generative AI, such as access “to materially nefarious information” related to weapons, violence, hate speech, obscene imagery, or ecological damage.  

The EU Artificial Intelligence Act specifically mandates AI developers of generative AI based on Large Language Models (LLMs) to comply with rigorous obligations prior to placing on the market such systems, including design specifications, information relating to training data, computational resources to train the model, estimated energy consumption, and compliance with copyright laws associated with harvesting of training data.  

AI regulations and risk management frameworks provide the basis for establishing ethical guidelines that developers ought to follow. They ensure that AI technologies are developed and deployed in a manner that respects human rights and societal values.

Ultimately embracing responsible AI regulations and risk management frameworks deliver positive business outcomes as there is “an economic incentive to getting AI and gen AI adoption right. Companies developing these systems may face consequences if the platforms they develop are not sufficiently polished – and a misstep can be costly. 

Major gen AI companies, for example, have lost significant market value when their platforms were found hallucinating (when AI generates false or illogical information). Public trust is essential for the widespread adoption of AI technologies, and AI laws can enhance public trust by ensuring that AI systems are developed and deployed ethically. 

You may also like…

Q&A: Evaluating the ROI of AI implementation

From diagrams to design: How AI transforms system design