Here are some key policy and regulations areas that we should expect in 2021 that will affect companies that develop and deploy AI:
- Every federal agency will be required to develop an AI regulatory plan
Although politics have changed in the US following the elections win by Joe Biden, the government issued AI regulatory guidance in November 2020. The guidance calls for AI regulation at the federal level and requires agencies to consider non-regulatory approaches. It encourages the use of voluntary frameworks. Some of these frameworks include the National Institute of Standards (NIST) and privacy frameworks. The regulations suggest more regulation in some high-risk areas concerning things such as privacy and safety. Technologies such as facial recognition and autonomous vehicles have for some time faced criticisms over privacy concerns, and these are the areas that such regulations are expected to address.
- Regulations will demand collaboration between leaders and AI experts
As AI investment continues increasing and the scale of deployment continues rising, the collaboration between leaders and AI experts must be streamlined to determine the AI use cases and deliver a return on investment. The regulations will seek to ensure the greatest challenges in AI deployment are avoided, and confidence is assured in the deployment. Furthermore, they will ensure that organizations have the right data and can support AI projects.
- FTC enforcement activity
The rise of AI initiatives will see an increase in scrutiny aimed at AI and its algorithms. FTC seeks to ensure transparent, fair, and explainable use of AI tools. There will be increased monitoring of data-driven AI algorithms. Machine learning, code, and data used to train algorithms will be streamlined to enhance fairness and superior decision-making. There have been talks about what is known as “algorithmic justice,” which aims to remedy serious harm caused by faulty inputs, discrimination, and incorrect conclusions. The FTC enforcement regulations will require greater AI transparency and companies to bear responsibility, carry out regular audits and impact assessment. Furthermore, it seeks to facilitate appropriate redress for unfair algorithms.
- NIST Trustworthy AI Standards
NIST has already released a plan that engages federal leadership on AI standards. The body will engage other federal players in developing AI standards, starting from supporting and carrying out research and development. In 2021, NIST will develop additional standards that will ensure that AI is explainable, has no bias, and is secure. Many stakeholders have accused AI of being biased on specific groups of people, such as people of color. It is now time for NIST to find solutions that will assure fairness in algorithms through fair training data.
In summary, 2021 is yet another year that agencies- both government and private- will ensure that AI is fair for all. AI governance will become a board-level issue that will enhance competitiveness based on its potential to improve governance and business operational process and innovation. The expansion of research funding and coordination will place the federal government in a more active role in AI research. After gaining momentum in the past year, this year may see the development of consequential AI policies that will have far-reaching consequences in this technology.