{"id":189,"date":"2024-01-11T11:14:31","date_gmt":"2024-01-11T15:14:31","guid":{"rendered":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/?p=189"},"modified":"2024-01-11T11:14:31","modified_gmt":"2024-01-11T15:14:31","slug":"the-ai-update-january-11-2024","status":"publish","type":"post","link":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/2024\/01\/11\/the-ai-update-january-11-2024\/","title":{"rendered":"The AI Update | January 11, 2024"},"content":{"rendered":"<p><img loading=\"lazy\" decoding=\"async\" class=\"alignleft wp-image-96 size-full\" src=\"http:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-content\/uploads\/sites\/63\/2023\/04\/DM-AI-Update-e1681141844877.png\" alt=\"\" width=\"150\" height=\"60\" \/><em>#HelloWorld. It\u2019s 2024 and we\u2026 are\u2026back. Lots to catch up on. AI legal developments worldwide show no signs of letting up, so here\u2019s our New Year\u2019s resolution: We\u2019re redoubling efforts to serve concise, focused guidance of immediate use to the legal and business community. Let\u2019s stay smart together. (<a href=\"mailto:AI-Update@duanemorris.com?subject=Subscribe%20to%20the%20mailing%20list%20&amp;body=Please%20add%20me%20to%20The%20AI%20Update%20list.\">Subscribe to the mailing list<\/a>\u00a0to receive future issues.)<\/em><\/p>\n<p><!--more--><\/p>\n<p><strong>All the news that\u2019s fit to model. <\/strong>The current wave of generative AI copyright litigation began in November 2022 with <a href=\"https:\/\/www.courtlistener.com\/docket\/65669506\/doe-1-v-github-inc\/\" target=\"_blank\" rel=\"noopener\"><em>Doe v. Github<\/em><\/a> in the Northern District of California. But it crested right after Christmas last month when the New York Times filed a <a href=\"https:\/\/www.nytimes.com\/2023\/12\/27\/business\/media\/new-york-times-open-ai-microsoft-lawsuit.html\" target=\"_blank\" rel=\"noopener\">many<\/a>&#8211;<a href=\"https:\/\/www.theverge.com\/2023\/12\/27\/24016212\/new-york-times-openai-microsoft-lawsuit-copyright-infringement\" target=\"_blank\" rel=\"noopener\">headline<\/a>&#8211;<a href=\"https:\/\/www.washingtonpost.com\/technology\/2023\/12\/27\/new-york-times-sues-openai-chatgpt\/\" target=\"_blank\" rel=\"noopener\">grabbing<\/a> suit in the Southern District of New York. The suit charges OpenAI and Microsoft with infringing copyrights in \u201cmillions\u201d of Times articles.<\/p>\n<p>The <a href=\"https:\/\/storage.courtlistener.com\/recap\/gov.uscourts.nysd.612697\/gov.uscourts.nysd.612697.1.0.pdf\" target=\"_blank\" rel=\"noopener\">complaint<\/a> raises legal questions that are in many ways similar to those in the dozen or so <a href=\"https:\/\/blogs.duanemorris.com\/artificialintelligence\/2023\/10\/25\/the-ai-update-october-25-2023\/#more-161\" target=\"_blank\" rel=\"noopener\">other<\/a> AI copyright suits filed throughout 2023\u2014for example, is it fair use to train a large language model on a content corpus that includes copyrighted works? But aside from the outsize media attention, here are three ways the New York Times\u2019 suit is different:<\/p>\n<ul>\n<li style=\"list-style-type: none\">\n<ul>\n<li>Unlike some of the earlier plaintiffs, the Times appears to have registered with the Copyright Office all the works at issue <em>before<\/em> the alleged infringement began. This would make the Times eligible for statutory damages ranging from $750 to $30,000 per work\u2014and the complaint attaches thousands of pages of charts listing \u201cover 3 million\u201d articles the Times owns.<\/li>\n<li>One part of the complaint focuses on a specialized \u201cretrieval augmented generation\u201d process. In general, RAG occurs at the time the user prompts a model for a response (not earlier in LLM training) and typically surfaces substantial chunks of text that are then run through the already-trained model. Unlike basic LLM training in general, this RAG process in particular further complicates some of the debate around whether it\u2019s fair use to train a model on copyrighted content\u2014because RAG occurs <em>after <\/em>training and typically homes in on fewer documents than are used to train the model in the first place.<\/li>\n<li>It\u2019s clear from attachments to the complaint (especially Exhibit J) that many of the challenged LLM outputs resulted from highly specific and purpose-driven user prompts. For instance, in some cases, the Times\u2019 representatives pasted in paragraphs from an article and asked the model to reproduce the remaining parts. This raises an important question: In such situations\u2014where an individual user intentionally seeks to induce substantial copying\u2014who in the end should be held accountable? Expect the allocation-of-responsibility question to be a key battleground as the case proceeds. In fact, only a few days ago, OpenAI publicly <a href=\"https:\/\/openai.com\/blog\/openai-and-journalism\" target=\"_blank\" rel=\"noopener\">responded<\/a> to the suit by alleging that the Times \u201cis not telling the full story\u201d; it \u201cseems they intentionally manipulated prompts\u2026 in order to get our model to regurgitate\u2026 our models don\u2019t typically behave that way.\u201d<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n<p><strong>But how much is that content <em>really <\/em>worth?<\/strong> As the Times\u2019 complaint acknowledges, the lawsuit is an outgrowth of content-licensing discussions with OpenAI that began in April 2023 and reached an apparent impasse sometime in December. One side benefit of the Times\u2019 filing suit: Those of us obsessed with content valuations for AI use were treated to behind-the-scenes reports about negotiations between news providers and major AI tech developers\u2014including some of the numbers exchanged. On January 4, <a href=\"https:\/\/www.theverge.com\/2024\/1\/4\/24025409\/openai-training-data-lowball-nyt-ai-copyright\" target=\"_blank\" rel=\"noopener\"><em>The Verge<\/em><\/a> had this handy recap: OpenAI has been offering \u201cbetween $1 million and $5 million a year to license copyrighted news articles to train its AI models\u201d while Apple \u201cis offering at least $50 million over a multiyear period,\u201d but apparently for a broader scope of use. And back in December, it was <a href=\"https:\/\/finance.yahoo.com\/news\/openai-axel-springer-ink-deal-145333092.html\" target=\"_blank\" rel=\"noopener\">reported<\/a> that OpenAI had signed a deal with Axel Springer, the German owner of web properties including <em>Politico<\/em> and <em>Business Insider<\/em>, valued at \u201ctens of millions of euros\u201d over three years.<\/p>\n<p><strong>Boards may soon need to speak up on AI. <\/strong>On the corporate governance front, in a pair of parallel January 3 notice letters, the SEC informed <a href=\"https:\/\/www.sec.gov\/files\/corpfin\/no-action\/14a-8\/aflciodisney010324-14a8.pdf\" target=\"_blank\" rel=\"noopener\">Disney<\/a> and <a href=\"https:\/\/www.sec.gov\/files\/corpfin\/no-action\/14a-8\/aflcioapple010324-14a8.pdf\" target=\"_blank\" rel=\"noopener\">Apple<\/a> they would have to allow an AI-focused shareholder proposal (pushed by the AFL-CIO) onto their respective annual meeting proxy statements. The proposal seeks a vote on whether each company should be required to \u201cprepare a transparency report on the company\u2019s use of Artificial Intelligence\u201d and to \u201cdisclose any ethical guidelines that the company has adopted regarding the company\u2019s use of AI technology.\u201d Interestingly, the SEC determined that these subjects \u201ctranscend ordinary business matters.\u201d Maybe one day, heck maybe soon, AI tools will be as entrenched in the enterprise as Microsoft Word. But not just yet, according to our chief securities regulator. Nonetheless, time to revisit those AI use policies\u2026.<\/p>\n<p><strong>Lest we forget our friends in Europe. <\/strong>We may have been away in December, but the European Union was not. On December 9, the European Parliament and Council reached \u201c<a href=\"https:\/\/www.consilium.europa.eu\/en\/press\/press-releases\/2023\/12\/09\/artificial-intelligence-act-council-and-parliament-strike-a-deal-on-the-first-worldwide-rules-for-ai\/\" target=\"_blank\" rel=\"noopener\">provisional agreement<\/a>\u201d on the EU AI Act, in the making for almost three years. This important, intricate legislation does not lend itself to The AI Update\u2019s punchy summarization and, in any event, technical details are still being hammered out, the final text has not been released, and formal adoption remains. But rest easy, we\u2019ll continue to keep close watch throughout 2024.<\/p>\n<p><strong>What we\u2019re reading. <\/strong>Before we go, here\u2019s a fun experiment: Can you induce a chatbot (in an artificial environment, of course) to make \u201cillegal trades based on insider information\u201d and then \u201cstrategically hide\u201d that conduct in its answers, even when said bot was previously \u201ctrained to be helpful, harmless, and honest\u201d? The recent answer from three safety researchers apparently is yes; their secret sauce was prompting the bot with simulated \u201chigh-pressure\u201d emails. Read the details <a href=\"https:\/\/arxiv.org\/pdf\/2311.07590.pdf\" target=\"_blank\" rel=\"noopener\">here<\/a>, while we hope with you that this merely is a useful safety-promoting exercise\u2014and not the dawn of GPT-Boesky.<\/p>\n<p><strong>What <em>should <\/em>we be following?<\/strong> Have suggestions for legal topics to cover in future editions? Please send them to <a href=\"mailto:AI-Update@duanemorris.com\">AI-Update@duanemorris.com<\/a>. We\u2019d love to hear from you and continue the conversation.<\/p>\n<p><strong><em>Editor-in-Chief<\/em><\/strong><strong>: <\/strong><a href=\"mailto:agoranin@duanemorris.com\">Alex Goranin<\/a><\/p>\n<p><strong><em>Deputy Editors<\/em><\/strong><strong>:<\/strong> <a href=\"mailto:mcmousley@duanemorris.com\">Matt Mousley<\/a> and <a href=\"mailto:tmarandola@duanemorris.com\">Tyler Marandola<\/a><\/p>\n<p><em>If you were forwarded this newsletter, <\/em><a href=\"mailto:AI-Update@duanemorris.com?subject=Subscribe%20to%20the%20mailing%20list%20&amp;body=Please%20add%20me%20to%20The%20AI%20Update%20list.\"><em>subscribe to the mailing list<\/em><\/a><em> to receive future issues.<\/em><\/p>\n","protected":false},"excerpt":{"rendered":"<p>#HelloWorld. It\u2019s 2024 and we\u2026 are\u2026back. Lots to catch up on. AI legal developments worldwide show no signs of letting up, so here\u2019s our New Year\u2019s resolution: We\u2019re redoubling efforts to serve concise, focused guidance of immediate use to the legal and business community. Let\u2019s stay smart together. (Subscribe to the mailing list\u00a0to receive future &hellip; <\/p>\n<p class=\"link-more\"><a href=\"https:\/\/blogs.duanemorris.com\/artificialintelligence\/2024\/01\/11\/the-ai-update-january-11-2024\/\" class=\"more-link\">Continue reading<span class=\"screen-reader-text\"> &#8220;The AI Update | January 11, 2024&#8221;<\/span><\/a><\/p>\n","protected":false},"author":6,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[27,15,141,16,31,140,17],"ppma_author":[5],"class_list":["post-189","post","type-post","status-publish","format-standard","hentry","category-general","tag-ai","tag-alex-goranin","tag-journalism","tag-matt-mousley","tag-openai","tag-publications","tag-tylermarandola"],"authors":[{"term_id":5,"user_id":6,"is_guest":0,"slug":"duanemorris3","display_name":"Duane Morris","avatar_url":"https:\/\/secure.gravatar.com\/avatar\/843ff6e7a8fe5fc92109b47a45f34b6cf0ea499e6e788db23456c838b0ae6747?s=96&d=blank&r=g","0":null,"1":"","2":"","3":"","4":"","5":"","6":"","7":"","8":""}],"_links":{"self":[{"href":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-json\/wp\/v2\/posts\/189","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-json\/wp\/v2\/comments?post=189"}],"version-history":[{"count":0,"href":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-json\/wp\/v2\/posts\/189\/revisions"}],"wp:attachment":[{"href":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-json\/wp\/v2\/media?parent=189"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-json\/wp\/v2\/categories?post=189"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-json\/wp\/v2\/tags?post=189"},{"taxonomy":"author","embeddable":true,"href":"https:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-json\/wp\/v2\/ppma_author?post=189"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}