TL;DR:
- Congress faces mounting pressure to regulate artificial intelligence (AI) and is considering a bill with bipartisan support.
- The American Data Privacy and Protection Act (ADPPA) aims to enhance data privacy rights while regulating emerging AI systems.
- Companies, including Microsoft and Salesforce, support the bill as it would help build trust in AI products.
- The bill addresses discriminatory AI output in business settings and prohibits data usage that discriminates based on factors like sex or race.
- Companies would need to evaluate their AI systems and allow government scrutiny to ensure compliance.
- The focus is on high-risk AI systems used in consequential decisions related to credit, housing, employment, healthcare, and insurance.
- Efforts are underway to refine the bill’s language and define “high-risk” settings and “consequential decisions.”
- Senate Majority Leader Chuck Schumer is developing an AI framework, but the ADPPA has had bipartisan support since last year.
- The House Energy and Commerce Committee, chaired by Rep. Cathy McMorris Rodgers, is interested in revisiting the bill.
- Passing a federal data privacy and security law is seen as crucial for providing certainty and safety in AI development.
Main AI News:
Here’s how the first wave of AI regulations from Congress could take shape, according to a potential bill that garnered bipartisan support in the previous session of Congress, led by Democrats.
The American Data Privacy and Protection Act (ADPPA), passed by the House Energy and Commerce Committee in 2022, aims to bolster data privacy rights while also serving as a crucial tool in the regulation of emerging AI systems. With strong bipartisan backing last year, the ADPPA continues to enjoy support from companies eager to establish trust in their AI products, as they recognize the need for a federal regulatory framework to achieve this objective.
BSA/Software Alliance, representing prominent companies such as Microsoft, Okta, and Salesforce, is actively collaborating with the committee to ensure the passage of a revised version of the bill this year, with hopes of securing approval in a full House vote. Craig Albright, BSA’s Vice President for U.S. Government Relations, emphasized the practicality and importance of the proposed legislation, stating, “We think this is a very digestible bite… an important element of guardrails for AI. And we think that this is doable this year.”
Addressing a key concern for BSA members, the previous version of the bill tackles the issue of discriminatory AI output in business contexts. This form of discrimination pertains to situations where an AI system, due to design flaws, faulty data, misinterpretation, or other reasons, generates recommendations that could lead to unfavorable outcomes, such as denying someone a mortgage loan.
It is distinct from the “bias” problem observed in generative AI systems like ChatGPT, where concerns revolve around politically skewed responses or deliberate avoidance of provocative replies.
The ADPPA seeks to combat such “discrimination” by prohibiting companies from collecting, transferring, or using data in ways that discriminate based on sex, race, disability, or other factors that limit individuals’ access to goods and services.
Companies would be required to evaluate their systems for such flaws and allow government scrutiny of these assessments. BSA also supports the notion of companies obtaining certification from a federal regulator once they adhere to these guidelines.
According to Albright, this regulatory approach aims to instill trust in AI within “high-risk” systems employed for making “consequential decisions.” He explained, “What we’re working with the committee to do is provisions that would establish what high-risk means, and it’s hooked to the phrase ‘consequential decisions.’ So, we’re talking about AI systems that are used in decisions about credit, housing, employment, health care, insurance, those kinds of decisions.”
To ensure practicality and effectiveness, ongoing efforts involve refining the language of the provisions and defining “high-risk” AI settings and “consequential decisions” more explicitly. Albright emphasized the need to update these provisions, stating, “What we are urging the committee to do is to update those provisions to make sure that they are workable and effective.”
While the ADPPA is not the sole legislative proposal in the AI realm, Senate Majority Leader Chuck Schumer, D-N.Y., has discussed the development of an AI framework that is still being shaped. However, BSA remains optimistic due to the bipartisan support garnered by last year’s bill and the positive responses received from lawmakers during discussions about reintroducing it this year.
Under the leadership of Rep. Cathy McMorris Rodgers, R-Wash., the House Energy and Commerce Committee has also shown interest in resuming efforts on the bill. The committee has conducted several hearings on data privacy, and an aide indicated that the ADPPA remains under consideration.
According to the aide, “The most important thing we can do to begin providing certainty and safety to the development of AI is passing a federal data privacy and security law.”
Conlcusion:
The potential implementation of the American Data Privacy and Protection Act (ADPPA) and the growing bipartisan support for regulating artificial intelligence (AI) signify significant developments in the market. These regulatory efforts demonstrate a proactive approach by Congress to address concerns surrounding AI, particularly in terms of data privacy and discriminatory output. Companies operating in the AI sector, including tech giants like Microsoft and Salesforce, can anticipate a more structured regulatory landscape that fosters trust in AI products.
Moreover, the focus on high-risk AI systems used in consequential decisions indicates a heightened emphasis on ensuring fairness and accountability. These regulatory measures, coupled with ongoing discussions on data privacy and security laws, signal a market environment that prioritizes responsible AI development and provides clearer guidelines for businesses to navigate the evolving landscape.
