Welcome to the issue of this AI week: Washington’s reportA joint company in Mintz and its government affiliate, ML Strategies. The acceleration of progress of artificial intelligence (“AI”) and the practical, legal and political problems created by AI have exponentially increased the interest of the federal government for AI and its implications. In these weekly reports, we hope to keep our customers and friends aware of this set of potential legislative, executive and regulatory activities focused on Washington.
In this issue, we discuss July 2023 Civil investigation request (“CID”) The Federal Trade Commission (“FTC” or “Commission”) was used for Chatgpt Developer Openai. Our main points to remember are:
- OPENAI CID concerns the question of whether OPENAI “has embarked on confidentiality or security practices of unfair or misleading data” or “engaged in unfair or misleading practices linked to the risk of damage to consumers, including The prejudice of reputation… ”
- By using the information collected from the CID, the Commission may alleviate that the conduct of Openai raped Article 5 of the FTC lawpossibly resulting in fines or a consent decree.
- For the FTC, this survey is an important and concrete step towards the realization of the Declared objective of the commission to apply its consumer protection powers to the AI field.
The request for a civil investigation of the FTC was used to the signals to open the intention to apply the current law at the AI
On July 13, 2023, the Washington Post published partially expounded Civil investigation request (“Cid” or “Openai Cid”) served by the Federal Trade Commission (“FTC”) to OPENAIThe Chatgpt developer. The OPENAI CID contains 49 requests for information related to business practices and requests 17 classes of documents. The declared subject of the FTC survey is to determine whether OPENAI “engaged in confidentiality or security practices of unfair or misleading data” or “engaged in unfair or misleading practices relating to the risk of damage to consumers, including reputation damages… ”
Like us discuss In previous newsletters, in the absence of clear regulatory guidelines for Congress, regulatory agencies have sought to use their existing authority to regulate the growing tide of AI products and services. Of these regulatory organizations, it is the FTC which was the most active In this area, Release non -binding commercial adviceco-signer joint declaration on biases in automated systems,(1) and pioneer of a new form of application called Algorithmic dispute.
As discussed in our newsletter on FTC incursions in the AI applicationThe current committee affirms that the statutory authority of the FTC under laws – including article 5 of the FTC law, the law on credit reports and the law on equality of credit opportunities – grants him the Ability to regulate the conduct of AI developers whose practices violate these laws. In the case of the OPENAI CID, the FTC asked the company’s information and documents to determine if there is one case to make so that the conduct of OPENAI has violated article 5 of the FTC law.(2)
The subject of the OPENAI CID
As indicated above, the Openai Cid cares about whether the company has committed one of the following alleged violations of the FTC Act.
- Mishandling user data: That OpenAi has engaged in data confidentiality practices or disheveled or misleading data.
- Cause consumers’ reputation damage: That Openai has engaged in unfair or deceptive practices which risked harm to consumers. The Commission’s objective is the question of the extent to which Openai products cause “reputation damages” to consumers by creating “deceitful, derogatory or harmful false declarations” on individuals.
The previous surveys of the FTC dealt with these two classes of potential violations of article 5. The Commission has concluded a certain number of consent decrees with large technological companies concerning their confidentiality and data security practices, in particular with Facebook in November 2011. The FTC also presented cases which have included allegations of reputation damage to consumers, including a 2002 online confidentiality case involving the pharmaceutical company Eli Lily allegedly disclosed “A list of emails of more than six hundred consumers using Prozac” without the consent of these customers.
What is notable in the Openai Cid is therefore not the potential violations of the FTC law that the commission investigates, but rather the role of the CID in the continuous effort of the FTC to assert its principal role IA regulator. While FTC commercial advice And Leadership declarations pointed out the Commission’s intention to regulate this space, the portion of the Openai Cid represents a significant and concrete step towards the achievement of this objective. Regardless of the fact that the FTC is able to build a case that open violations committed section 5, and if such a hypothetical case has merit, the outcome of this survey could have important implications to determine the scope of the existing authority of the FTC to regulate the developers of AI.
Contents of the Openai Cid
To determine whether OPENAI “has engaged in confidentiality or security practices of unfair or misleading data”, the FTC asked that the company provides the following information concerning its business practices.
- The data used to train important OPENAI (“OpenAi” models) models, the sources of said data, the content categories included in this data and all the policies and procedures related to the verification and selection of said data.
- Internal policies concerning the surveillance of human examiners of the OpenAi models.
- The procedures followed to assess the safety of products manufactured with OPENAI models before its launch. Process surrounding recycling and refining of OpenAi models.
- Process to prevent personal information from being included in the training data of OPENAI models.
- Measures taken to mitigate the exposure of personal information from users via integration and API plug-ins.
- OPENAI Management of a March 2023 Security incident in which the sensitive personal data of users have been temporarily disclosed.
To test if OPENAI “has engaged in unfair or misleading practices linked to the risk of harm to consumers, including reputation damage”, the FTC asked for the following information:
- Policies and procedures related to the generation of OpenAi’s large -language model declarations on individuals, including the relationship management processes that the models “have done … false, derogatory, derogatory or harmful declarations” on individuals.
- Measures taken to test and respond to the capacity of products manufactured with OPENAI models to generate “declarations on real people who are false, deceptive or derogatory”.
- Process to determine if an individual is a “public figure” and if this determination has an impact on the treatment with the model of great language of this individual.
Conclusion: A model for future APTC application actions on AI?
By serving a CID to an AI developer to Leader Openai, the FTC expressed its point of view according to which it can, under the current law, of the police of certain aspects of the evolution of the AI. THE Chatgpt release In November 2022, launched the current wave of interest in the regulation of AI observed in the world. The CEO of Openai, Sam Altman, was a visible presence in the rooms of the Congress, as he did Called to AI regulation in several conference hearings. The FTC surveys are normally confidential, so that the details surrounding the course of this CID and the FTC assessment of these problems with Openai will be difficult to obtain in real time. We will continue to monitor FTC initiatives in this area.