While the FTC appreciates the enormous potential of AI to improve welfare and productivity, it also is mindful of the risks.

In a blog, the FTC emphasized its actions such as bringing suits against businesses that violate laws relating to AI and automated decision making.

In particular, it cited laws such as the Fair Credit Reporting Act (FCRA), enacted in 1970, and the Equal Credit Opportunity Act (ECOA), enacted in 1974. The Commission said it used the FTC Act to avert injury to consumers from AI.

The consumer watchdog also codified its principles for the protection of consumers from AI and algorithms. (ftc.gov)

AI and the FTC’s five-point mantra

“The FTC’s law enforcement actions, studies, and guidance emphasize that the use of AI tools should be transparent, explainable, fair, and empirically sound while fostering accountability,” the watchdog said. “We believe that our experience, as well as existing laws, can offer important lessons about how companies can manage the consumer protection risks of AI and algorithms.”

Transparent:

Companies using consumer data in AI applications should clearly inform consumers of the fact. They should also clarify the manner of collection of the data and how they process it. Companies should strictly comply with the FCRA in this regard. Further, AI should not mislead consumers (Example: Dating service Ashley Madison’s use of fake “engager profiles” of attractive mates)

Explainable:

You must know what data is used in your model and how that data is used to arrive at a decision. And you must be able to explain that to the consumer. (Example: Automated denial of credit based on credit score should be fully explainable by actual data)

Fair:  

There should be no discrimination against protected classes. The FTC enforces ECOA, which prohibits credit discrimination based on race, color, religion, national origin, sex, marital status, age, or because a person receives public assistance. Is there any bias in the outcomes from the AI model? (Example: A company making credit decisions based on consumers’ Zip Codes, resulting in a “disparate impact” on particular ethnic groups). Companies should also ensure that any consumers aggrieved with the data should have the opportunity to correct it.

Empirically sound:

If you are a consumer reporting agency that others use to decide on credit, employment, insurance, housing, government benefits, check-cashing or similar transactions, you are responsible under the FCRA. Ensure that your data is accurate and current. (Example: RealPage, Inc was slapped with a penalty of $3 million for rendering inaccurate information on applicants’ criminal information to landlords and property managers)

Make sure that your AI models are validated and revalidated to ensure that they work as intended. They should not also illegally discriminate.

Accountable:  

In a hark back to its 2016 Big Data report, the Commission warned companies that big data analytics could result in bias or other harm to consumers. It suggested that outside tools and services could audit the algo. It also asked that companies prevent any kind of unauthorized use of the algorithm.

Related Story:   Artificial Intelligence: The Pentagon Adopts Five AI Principles