![\"\"](\"http:\/\/blogs.duanemorris.com\/artificialintelligence\/wp-content\/uploads\/sites\/63\/2023\/04\/DM-AI-Update-e1681141844877.png\")
<\/p>\n<\/p>\n
#HelloWorld. Fall begins, and the Writers Guild strike ends. In this issue, we look at what that means for AI in Hollywood. We also run through a dizzying series of self-regulating steps AI tech players have undertaken. As the smell of pumpkin spice fills the air, let\u2019s stay smart together. (<\/em>Subscribe to the mailing list<\/em><\/a>\u00a0to receive future issues.)<\/em><\/p>\n AI and the Writers Guild.<\/strong> Screenwriters ended a nearly five-month strike with a tentative new agreement good through mid-2026. The Minimum Basic Agreement<\/a> (MBA) includes multiple AI-centric provisions. The AI-related highlights:<\/p>\n One of the mostly debated issues (also at play in many recent copyright<\/a> litigations<\/a>) is whether writers\u2019 works can be used to train generative AI models. The MBA saves that fight for another day: Writers reserve the right to \u201cassert that exploitation of writers\u2019 material to train AI is prohibited by\u201d the Minimum Basic Agreement \u201cor other law.\u201d<\/p>\n Industry continues to self-regulate\u2014and pay.<\/strong> One theme emerging from the Writers Guild compromise\u2014disclosure of AI usage\u2014has echoes in the AI industry at large. Social media sensation TikTok, for instance, recently announced<\/a> a new collection of tools and technologies for labeling AI material posted by creators to the platform. One tool enables creators \u201cto easily inform their community when they post AI-generated content.\u201d Another tool under development would automatically detect content \u201cedited or created with AI\u201d and programmatically affix an \u201cAI-generated\u201d label to it.<\/p>\n Other platforms are moving beyond labeling to compensation. As previously reported, some owners of large (and therefore valuable) datasets have negotiated licensing deals for AI training directly with AI model developers. The separate deals OpenAI made this summer with Shutterstock<\/a> and the Associated Press<\/a> are two cases in point.<\/p>\n But what if you\u2019re an individual artist without a huge aggregated portfolio of works to license? Enter Adobe. While the software company maintains that it has the legal right to train its Firefly image-synthesizing AI model on works uploaded to its platform, Adobe has started paying \u201cbonuses\u201d<\/a> to artists whose images were used in training. The payout is a function of (a) the number of images submitted and (b) the number of times someone licensed those images in the preceding 12 months. Unfortunately, Adobe hasn\u2019t disclosed how big the bonuses paid to date have been, but an educated guess places them closer to the $1 to $100 end of the spectrum than the $1,000 to $10,000 end. Valuation of individual contributions to a training dataset is one of the biggest unsolved issues in this space.<\/p>\n Another red team for OpenAI.<\/strong> Last week OpenAI announced the start of an OpenAI Red Teaming Network<\/a>. \u201cRed teaming\u201d is not new; it refers to the process of testing systems for weaknesses by actively trying to find vulnerabilities. Cybersecurity defense against malicious hackers is a classic use case. In the generative AI context, think of someone attempting to \u201ctrick\u201d an AI model through various prompt variations into giving harmful advice, spreading misinformation, disclosing proprietary data, or leaking private personal information. According to OpenAI\u2019s blog, the company\u2019s Red Teaming Network will have experts recruited from a broad range of subjects involved in \u201crigorously evaluating\u201d future iterations of OpenAI products. The list of desired domain expertise reads like a college curriculum: cognitive science, biology, computer science, political science, persuasion, anthropology, finance, healthcare, biometrics, psychology, law, chemistry, and on and on.<\/p>\n What we\u2019re reading: <\/strong>Speaking of college reminiscences, we came across an interesting paper<\/a> from a University of Pennsylvania and Google Research team, asking the question: Can a language model be trained to direct a Dungeons & Dragons game, responding \u201cas the player who runs the game\u2014i.e., the Dungeon Master\u201d? Interestingly, the team\u2019s answer encapsulates the current state of LLMs: Models can generate \u201cevocative, in-character text\u201d to serve as \u201cinspiration\u201d for human<\/strong> Dungeon Masters, but they cannot quite track \u201cthe full state\u201d of the game for long enough to enable \u201cfully autonomous\u201d DMs. Humans live to fight another day, no dodecahedral dice roll required\u2026yet.<\/p>\n What should <\/em>we be following?<\/strong> Have suggestions for legal topics to cover in future editions? Please send them to AI-Update@duanemorris.com<\/a>. We\u2019d love to hear from you and continue the conversation.<\/p>\n Editor-in-Chief<\/em><\/strong>: <\/strong>Alex Goranin<\/a><\/p>\n Deputy Editors<\/em><\/strong>:<\/strong> Matt Mousley<\/a> and Tyler Marandola<\/a><\/p>\n\n
\n