The European Union is gearing up to regulate AI, but what is the U.S. doing?

• There are new Federal algorithmic transparency bills being filed:
  • The Algorithmic Accountability Act of 2022, introduced by Senator Ron Wyden (D-OR), Senator Booker (D-NJ), and Representative Clarke (D-NY), would require an impact assessment for automated processing and require the Federal Trade Commission to provide structured guidelines.
• There are AI specific laws dealing with Artificial Intelligence:
  • If you use AI to make hiring decisions in NYC, you must provide the applicants with notice that an automated employment decision tool will be used in connection with their assessment or evaluation, the job qualifications and characteristics that such an automated employment decision tool will use in the assessment of such candidate or employee, and information about the type of data they will collect. You will also need to conduct a bias audit.
  • There is a similar law in Illinois.
• New U.S. privacy laws care about Artificial Intelligence:
  • CPRA contemplates regulations to provide an opt out and meaningful information about the logic involved in those decision-making processes, as well as a description of the likely outcome of the process with respect to the consumer for profiling (automated processing of personal information to analyze or predict aspects concerning that natural person’s performance at work, economic situation, health, personal preference, etc.)
  • CPA and CDPA require an opt out and a DPIA for any profiling that is in furtherance of decisions that “produce legal or significantly similar effects” (Hi Art 22 GDPR)
• The old sectoral U.S. privacy laws care about Artificial Intelligence
  • With the increase in use of AI in hiring and employment decisions, the FCRA is getting new attention.
• The FTC cares about Artificial Intelligence
  • In a 2020 paper, the FTC issued informal guidance outlining principles and best practices surrounding transparency, explainability, bias and robust data models.
  • In 2021, the FTC in Everalbum required Everalbum to delete data, models and algorithms that it had developed by using photos and videos uploaded by its users without their express consent.
  • In 2020, the FTC required WW/Kurbo to delete or destroy any “Affected Work Product” (i.e., “models or algorithms developed in whole or in part using Personal Information Collected from Children”) and submit a sworn statement under penalty of perjury confirming such deletion or destruction in order to prevent secondary use of the data, as it was acquired without proper consent.

[View source.]