OPEN AI (FTC 2023)

20 March 2023

Dear Friends

In 2019, many countries around the world, including the United States, committed to the development of human-centric and trustworthy AI. Yet less than a few years on, we appear to be approaching a tipping point with the release of Generative AI techniques, which are neither human-centric nor trustworthy. 

These systems produce results that cannot be replicated or proven. They fabricate and hallucinate. They describe how to commit terrorist acts, how to assassinate political leaders, and how to conceal child abuse.  GPT-4 has the ability to undertake mass surveillance at scale, combining the ability to ingest images, link to identities, and develop comprehensive profiles.

As this industry has rapidly evolved so too has the secrecy surrounding the products. The latest technical paper on GPT-4 provides little information about the training data, the number of parameters, or the assessment methods. A fundamental requirement in all emerging AI policy frameworks – an independent impact assessment prior to deployment – was never undertaken. 

Many leading AI experts, including many companies themselves, have called for regulation. Yet there is little effort in the United States today to develop regulatory responses even as countries around the world race to establish legal safeguards. 

The present course cannot be sustained. The public needs more information about the impact of artificial intelligence. Independent experts need the opportunity to interrogate these models. Laws should be enacted to promote algorithmic transparency and counter algorithmic bias. There should be a national commission established to assess the impact of AI on American Society, to better understand the benefits as well as the risks.

This week the Center for AI and Digital Policy, joined by others, will file a complaint with the Federal Trade Commission, calling for an investigation of Open AI and the product chatGPT.  We believe the FTC has the authority to act in this matter and is uniquely positioned as the lead consumer protection agency in the United States to address this emerging challenge. We will ask the FTC to establish a moratorium on the release of further commercial versions of GPT until appropriate safeguards are established. We will simultaneously petition the FTC to undertake a rulemaking for the regulation of the generative AI industry. 

We favor growth and innovation. We recognize a wide range of opportunities and benefits that AI may provide. But unless we are able to maintain control of these systems, we will be unable to manage the risk that will result or the catastrophic outcomes that may emerge. We are asking the FTC to “hit the pause button” so that there is an opportunity for our institutions, our laws, and our society to catch up. We need to assert agency over the technologies we create before we lose control. 

Merve Hickock and Marc Rotenberg

For the CAIDP 

As we move forward with the FTC Complaint and Petition, we welcome your suggestions for points to make and issues to raise with the FTC. Most helpful for us are (1) accurate, authoritative descriptions of risks arising from the use of GPT, and (2) expert opinions of risks arising from the use of GPT, and (3) examples of how GPT violates the specific guidelines that the Federal Trade Commission has established for the marketing and advertising of AI products and services. 

Among the topics we have identified so far:

• Enhanced risk to cybersecurity
• Enhanced risk to data protection and privacy
• Enhanced risk to children's safety
• Failure to conduct independent risk assessment prior to deployment
• Failure to establish independent risk assessment throughout AI lifecycle
• Failure to accurately describe data source
• Failure to disclose data collection practices regarding users
• False advertising regarding reliability
• Lack of transparency in outputs produced
• Replication of bias in protected categories

Please note that we make evidence-based arguments and cite to published work. We will not be able to include general policy arguments, unsupported claims, or rhetorical statements.

Thank you for your assistance!

Over the last several years, the FTC has issued several reports and policy guidelines concerning marketing and advertising of AI-relate products and services. We believe that OpenAI should be required by the FTC to comply with these guidelines.

When you talk about AI in your advertising, the FTC may be wondering, among other things:

• Are you exaggerating what your AI product can do? 
• Are you promising that your AI product does something better than a non-AI product?
• Are you aware of the risks?
• Does the product actually use AI at all?

[More information]

In 2021, the FTC warned that advances in Artificial Intelligence  "has highlighted how apparently “neutral” technology can produce troubling outcomes – including discrimination by race or other legally protected classes." The FTC explained it has decades of experience enforcing three laws important to developers and users of AI:

• Section 5 of the FTC Act
• Fair Credit Reporting Act
• Equal Credit Opportunity Act

The FTC said its recent work on AI – coupled with FTC enforcement actions – offers important lessons on using AI truthfully, fairly, and equitably.

• Start with the right foundation
• Watch out for discriminatory outcomes
• Embrace transparency and independenc
• Don’t exaggerate what your algorithm can do or whether it can deliver fair or unbiased results
• Tell the truth about how you use data
• Do more good than harm
• Hold yourself accountable – or be ready for the FTC to do it for you

[More information]

" . . . we at the FTC have long experience dealing with the challenges presented by the use of data and algorithms to make decisions about consumers. Over the years, the FTC has brought many cases alleging violations of the laws we enforce involving AI and automated decision-making, and have investigated numerous companies in this space.

"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. 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."

• Be transparent.
• • Don’t deceive consumers about how you use automated tools
  • Be transparent when collecting sensitive data
  • If you make automated decisions based on information from a third-party vendor, you may be required to provide the consumer with an “adverse action” notice
• Explain your decision to the consumer.
• • If you deny consumers something of value based on algorithmic decision-making, explain why
  • If you use algorithms to assign risk scores to consumers, also disclose the key factors that affected the score, rank ordered for importance
  • If you might change the terms of a deal based on automated tools, make sure to tell consumers.
• Ensure that your decisions are fair.
• • Don’t discriminate based on protected classes.
  • Focus on inputs, but also on outcomes
  • Give consumers access and an opportunity to correct information used to make decisions about them
• Ensure that your data and models are robust and empirically sound.
• Hold yourself accountable for compliance, ethics, fairness, and nondiscrimination.

FTC Report Warns About Using Artificial Intelligence to Combat Online Problems

Agency Concerned with AI Harms Such As Inaccuracy, Bias, Discrimination, and Commercial Surveillance Creep (June 16, 2022)

Today the Federal Trade Commission issued a report to Congress warning about using artificial intelligence (AI) to combat online problems and urging policymakers to exercise “great caution” about relying on it as a policy solution. The use of AI, particularly by big tech platforms and other companies, comes with limitations and problems of its own. The report outlines significant concerns that AI tools can be inaccurate, biased, and discriminatory by design and incentivize relying on increasingly invasive forms of commercial surveillance.