![\"music](\"https:\/\/www.digitalmusicnews.com\/wp-content\/uploads\/2024\/12\/Anthropic-Injunction-Opposition-1024x546.jpg\" "\\"music\\"")
<\/p>\n
Anthropic has doubled down on its opposition to major music publishers\u2019 sought injunction. Photo Credit: Igor Omilaev<\/p>\n<\/div>\n
Music Anthropic is doubling down on its opposition to a preliminary injunction push from major music publishers, which say the AI giant\u2019s Claude chatbot has infringed on their protected works. <\/strong><\/h2>\nAmazon-backed<\/a> Anthropic just recently refiled its opposition to the plaintiff publishers\u2019 desired preliminary injunction, on the heels of a court order granting and denying in part jointly sought redactions.<\/p>\nNow, those redactions (impacting a relatively small portion of Anthropic\u2019s argument, albeit including interesting details like how much it would cost to comply with the injunctions) are reflected in the appropriate documents.<\/p>\n
For a bit of quick background, it\u2019s now been more than a year since Concord, Universal Music, and others first requested said relief<\/a>. One of the desired injunctions pertains to removing the plaintiffs\u2019 protected works from the generative AI system\u2019s training data, with the other calling for an order blocking protected lyrics from appearing in Claude outputs.<\/p>\nAt the top level, Anthropic in opposing the injunctions largely reiterated longstanding arguments \u2013 chief among them that using copyrighted works to train LLMs constitutes fair use.<\/h4>\n
\u201cWhile Anthropic is confident that using copyrighted content as training data for an LLM is a fair use under the law\u2014meaning that it is not infringement at all\u2014there is no basis to conclude that money damages would not make Plaintiffs whole if they ultimately prevailed on the merits,\u201d the AI platform indicated.<\/p>\n
\u201cThe Court, accordingly, should not stretch on this underdeveloped record to get ahead of the other cases where the analogous substantive copyright issue will be adjudicated on a full summary judgment or trial record,\u201d Anthropic continued.<\/p>\n
Unsurprisingly, Anthropic also reiterated at length the volume of data behind Clyde \u2013 the chatbot \u201clearns the patterns of language from trillions of tiny textual data points\u201d \u2013 and explored the training specifics, probably including \u201csome copyrighted works,\u201d at hand.<\/p>\n
Adjacent to the point, the AI developer stressed the now-common transformative argument. Utilizing \u201ccopyrighted song lyrics as part of a multi-trillion token dataset to train a generative AI model about the world and how language works is the very definition of \u2018transformative\u2019 under the fair use doctrine,\u201d per the defendant.<\/p>\n
Furthermore, certain cited research details \u201cpreceded the commercial release of Claude by nearly a year,\u201d according to the legal text. And since then, Anthropic has purportedly added \u201ca broad array of safeguards to prevent reproduction of copyrighted works from occurring in Claude outputs.\u201d<\/h4>\n
All told, those safeguards mean there\u2019s \u201cno reasonable expectation that the complained-of activity is likely to recur,\u201d according to Anthropic, which also refuted allegations of ongoing market and licensing harm stemming from the purported infringement.<\/p>\n
Lastly, in a separate declaration in support of the injunction opposition, Anthropic co-founder Jared Kaplan rattled off<\/a> his credentials and explained at length Claude\u2019s training particulars.<\/p>\nHeading into the new year, it\u2019ll be worth closely monitoring the status of the injunction battle \u2013 not to mention the overarching infringement dispute, a substantial portion of which might be tossed in the not-so-distant future<\/a>.<\/p>\n<\/section><\/div>\n
Now, those redactions (impacting a relatively small portion of Anthropic\u2019s argument, albeit including interesting details like how much it would cost to comply with the injunctions) are reflected in the appropriate documents.<\/p>\n
For a bit of quick background, it\u2019s now been more than a year since Concord, Universal Music, and others first requested said relief<\/a>. One of the desired injunctions pertains to removing the plaintiffs\u2019 protected works from the generative AI system\u2019s training data, with the other calling for an order blocking protected lyrics from appearing in Claude outputs.<\/p>\n \u201cWhile Anthropic is confident that using copyrighted content as training data for an LLM is a fair use under the law\u2014meaning that it is not infringement at all\u2014there is no basis to conclude that money damages would not make Plaintiffs whole if they ultimately prevailed on the merits,\u201d the AI platform indicated.<\/p>\n \u201cThe Court, accordingly, should not stretch on this underdeveloped record to get ahead of the other cases where the analogous substantive copyright issue will be adjudicated on a full summary judgment or trial record,\u201d Anthropic continued.<\/p>\n Unsurprisingly, Anthropic also reiterated at length the volume of data behind Clyde \u2013 the chatbot \u201clearns the patterns of language from trillions of tiny textual data points\u201d \u2013 and explored the training specifics, probably including \u201csome copyrighted works,\u201d at hand.<\/p>\n Adjacent to the point, the AI developer stressed the now-common transformative argument. Utilizing \u201ccopyrighted song lyrics as part of a multi-trillion token dataset to train a generative AI model about the world and how language works is the very definition of \u2018transformative\u2019 under the fair use doctrine,\u201d per the defendant.<\/p>\n All told, those safeguards mean there\u2019s \u201cno reasonable expectation that the complained-of activity is likely to recur,\u201d according to Anthropic, which also refuted allegations of ongoing market and licensing harm stemming from the purported infringement.<\/p>\n Lastly, in a separate declaration in support of the injunction opposition, Anthropic co-founder Jared Kaplan rattled off<\/a> his credentials and explained at length Claude\u2019s training particulars.<\/p>\n Heading into the new year, it\u2019ll be worth closely monitoring the status of the injunction battle \u2013 not to mention the overarching infringement dispute, a substantial portion of which might be tossed in the not-so-distant future<\/a>.<\/p>\n<\/section><\/div>\nAt the top level, Anthropic in opposing the injunctions largely reiterated longstanding arguments \u2013 chief among them that using copyrighted works to train LLMs constitutes fair use.<\/h4>\n
Furthermore, certain cited research details \u201cpreceded the commercial release of Claude by nearly a year,\u201d according to the legal text. And since then, Anthropic has purportedly added \u201ca broad array of safeguards to prevent reproduction of copyrighted works from occurring in Claude outputs.\u201d<\/h4>\n