![Microsoft pledges legal protection for AI-generated copyright breaches Microsoft pledges legal protection for AI-generated copyright breaches](https://i0.wp.com/www.ft.com/__origami/service/image/v2/images/raw/https://d1e00ek4ebabms.cloudfront.net/production/e1bd31ea-514a-4070-9cd0-ee3bc8e1eebf.jpg?fit=%2C&ssl=1)
Receive free AI updates
we will send you myFT Daily Digest Email summary of the latest information artificial intelligence Every morning there is news.
Microsoft has pledged to take legal responsibility for any copyright infringements it makes to businesses on material generated by artificial intelligence software in Word, PowerPoint and its coding tools, amid growing concerns over possible conflicts with content owners .
The U.S. tech giant pledged on Thursday to pay any legal fees for business customers sued for using artificial intelligence tools or any output.
Microsoft will protect paying customers of its $19-a-month GitHub Copilot, which uses generative artificial intelligence to create computer code, and Microsoft 365 Copilot, which applies artificial intelligence to products like Word, Teams and PowerPoint. 365 Copilot is still being tested by a small number of businesses.
Ilanah Fhima, professor of intellectual property law at University College London, said: “This opens up the market and makes the software more usable because it eliminates the need for businesses to know that they have that guarantee. one of the hurdles.”
The debate surrounding AI generation and copyright has led stock image providers such as Getty Images to file lawsuits against AI companies. Artists, singers, media companies and publishers allege that copyrighted material was used to train large language models without consent or compensation.
Adobe made a similar pledge in June to compensate users of its Firefly AI tool. The move by some of the world’s largest software developers is aimed at reassuring paying users amid growing concerns about the widespread use of generative artificial intelligence and its potential to produce passages of text or images that reproduce copyrighted sources the technology was trained on.
Hossein Nowbar, Microsoft’s general counsel for corporate legal affairs, acknowledged that clients are concerned about the risk of intellectual property infringement claims if the output produced by generative AI is used.
“This is understandable, given that writers and artists have recently been publicly asking how their own work can be used in conjunction with AI models and services,” he said in a blog post on Thursday.
“If you are challenged for copyright issues, we accept responsibility for the potential legal exposure involved,” he added.
If paying customers face legal problems for using “our co-pilot,” “we should treat that as our problem, not our customers’ problem,” Noba said.
If a third party sues a business customer for its use of Copilot or the output generated, Microsoft will defend the customer and “pay any adverse judgment or settlement amount resulting from the lawsuit, as long as the customer uses the guardrails and content we have built filters into in our products,” the blog post said.
Microsoft said its “guardrails” include content filters and detection of potentially infringing third-party content.
UCL’s Fhima emphasized that many of the controversies, laws and legal precedents surrounding AI and copyright are still developing, so they may not be costly to Microsoft.
“Technological developments are in the public interest, but strict copyright rights are not always enforced,” she added. “The risk Microsoft is taking is likely a calculated risk.”
Svlook