Voglio sapere a riguardo di…

6616

Federal Register / Vol. 86, No. 13 / Friday, January 22, 2021 / Notices
jbell on DSKJLSW7X2PROD with NOTICES

notice of proposed rulemaking or an advance notice of proposed rulemaking;
b. On agencies own comment submission forms, if agencies have them;
c. Within any automatic emails that agencies send acknowledging receipt of a comment;
d. On any part of agencies websites that describe their rulemaking process or within any rules on rulemakings they may have, as described in Recommendation 20201, Rules on Rulemakings; and e. Within any notices of public meetings pertaining to a rule.
3. The General Services Administrations eRulemaking Program Management Office should work with agencies that participate in Regulations.gov to include or refer to the notifications described in Paragraph 1 within any automated emails Regulations.gov sends acknowledging receipt of a comment.
4. If a submitter notifies an agency that the submitter inadvertently included protected material in the submitters comment, the agency should act as promptly as possible to determine whether such material warrants withholding from the public rulemaking docket and, if so, withhold it from the public rulemaking docket, or, if already disclosed, remove it from the public rulemaking docket.
If an agency determines that such material does not qualify as protected, it should promptly notify the submitter of this finding with a brief statement of reasons.
5. Agencies should allow third parties to request that protected material pertaining to themselves or a dependent be removed from the public rulemaking docket. Agencies should review such requests and, upon determining that the material subject to the request qualifies as protected material, should remove it from the public rulemaking docket as promptly as possible. If an agency determines that the material does not qualify as protected, it should promptly notify the requestor of this finding with a brief statement of reasons.
Recommendations for Agencies That Screen Comments for Protected Material Before Publication in the Public Rulemaking Docket 6. Agencies that screen comments for protected material before publication in the public rulemaking docket, either as required by law or as a matter of discretion, should redact the protected material and publish the rest of the comment. Redaction should be thorough enough to prevent the public from discerning the redacted material, but not so broad as to prevent the public from viewing non-protected material.
7. If redaction is not feasible within a comment, agencies should consider presenting the data in a summarized form.
8. If redaction is not feasible across multiple, similar comments, agencies should consider presenting any related information in an aggregated form. Agencies should work with data science experts and others in relevant disciplines to ensure that aggregation is thorough enough to prevent someone from disaggregating the information.
9. If the approaches identified in Paragraphs 68 would still permit a member of the public to identify protected material,
VerDate Sep<11>2014

19:27 Jan 21, 2021

Jkt 253001

agencies should withhold the comment in its entirety. When doing so, they should describe the withheld material for the public in as much detail as possible without compromising its confidentiality.
10. When deciding whether and how to redact, aggregate, or withhold protected material, agencies should explore using artificial intelligence-based tools to aid in identifying protected material. Agencies should consult with private sector experts and technology-focused agencies, such as the General Services Administrations Technology Transformation Service and the Office of Management and Budgets United States Digital Service, to determine which tools are most appropriate and how they can best be deployed given the agencies resources.
Recommendations for Agencies That Offer Assurances of Protection From Disclosure of Confidential Commercial Information 11. Agencies that offer assurances of protection from disclosure of confidential commercial information should decide how they will offer such assurances. Agencies can choose to inform submitters, directly upon submission, that they will withhold confidential commercial information from the public rulemaking docket; post a general notice informing submitters that confidential commercial information will be withheld from the public rulemaking docket; or both.
12. Such agencies should adopt policies to help them identify such information.
Agencies should consider including the following, either in tandem or as alternatives, as part of their policies, including within any rules on rulemakings they may have, as described in Recommendation 20201, Rules on Rulemakings:
a. Instructing submitters to identify clearly that the document contains confidential commercial information;
b. Instructing submitters to flag the particular text within the document that constitutes confidential commercial information; and c. Instructing submitters to submit both redacted and unredacted versions of a comment that contains confidential commercial information.
Administrative Conference Statement 20
Agency Use of Artificial Intelligence Adopted December 16, 2020
Artificial intelligence AI techniques are changing how government agencies do their work.1 Advances in AI hold out the promise 1 There is no universally accepted definition of artificial intelligence, and the rapid state of evolution in the field, as well as the proliferation of use cases, makes coalescing around any such definition difficult. See, e.g., John S. McCain National Defense Authorization Act for Fiscal Year 2019, Public Law 115232, 238g, 132 Stat. 1636, 169798 2018 using one definition of AI; Natl Inst. of Standards & Tech., U.S. Leadership in AI:
A Plan for Federal Engagement in Developing Technical Standards and Related Tools 78 Aug.
9, 2019 offering a different definition of AI.
Generally speaking, AI systems tend to have characteristics such as the ability to learn to solve complex problems, make predictions, or undertake tasks that heretofore have relied on human decision
PO 00000

Frm 00005

Fmt 4703

Sfmt 4703

of lowering the cost of completing government tasks and improving the quality, consistency, and predictability of agencies decisions. But agencies uses of AI also raise concerns about the full or partial displacement of human decision making and discretion.
Consistent with its statutory mission to promote efficiency, participation, and fairness in administrative processes,2 the Administrative Conference offers this Statement to identify issues agencies should consider when adopting or modifying AI
systems and developing practices and procedures for their use and regular monitoring. The Statement draws on a pair of reports commissioned by the Administrative Conference,3 as well as the input of AI experts from government, academia, and the private sector some ACUS
members provided at meetings of the ad hoc committee of the Administrative Conference that proposed this Statement.
The issues addressed in this Statement implicate matters involving law, policy, finances, human resources, and technology.
To minimize the risk of unforeseen problems involving an AI system, agencies should, throughout an AI systems lifespan, solicit input about the system from the offices that oversee these matters. Agencies should also keep in mind the need for public trust in their practices and procedures for use and regular monitoring of AI technologies.
1. Transparency Agencies efforts to ensure transparency in connection with their AI systems can serve many valuable goals. When agencies set up processes to ensure transparency in their AI
systems, they should consider publicly identifying the processes goals and the rationales behind them. For example, agencies might prioritize transparency in the service of legitimizing its AI systems, facilitating internal or external review of its AI-based decision making, or coordinating its AI-based activities. Different AI systems are likely to satisfy some transparency goals more than others. When possible, agencies should use metrics to measure the performance of their AI-transparency processes.
In setting transparency goals, agencies should consider to whom they should be making or intervention. There are many illustrative examples of AI that can help frame the issue for the purpose of this Statement. They include, but are not limited to, AI assistants, computer vision systems, biomedical research, unmanned vehicle systems, advanced game-playing software, and facial recognition systems as well as application of AI in both information technology and operational technology.
2 See 5 U.S.C. 591.
3 David Freeman Engstrom, Daniel E. Ho, Catherine M. Sharkey, & Mariano-Florentino Cuellar, Government by Algorithm: Artificial Intelligence in Federal Administrative Agencies Feb. 2020 report to the Admin. Conf. of the U.S., https www.acus.gov/report/government-algorithmartificial-intelligence-federal-administrativeagencies; Cary Coglianese, A Framework for Governmental Use of Machine Learning Dec. 8, 2020 report to the Admin. Conf. of the U.S., https www.acus.gov/report/frameworkgovernmental-use-machine-learning-final-report.

E:FRFM22JAN1.SGM

22JAN1