The issue of just how vital copyrighted data may or may not be to AI training in legal tech was recently thrown into sharp focus in a suit Thomson Reuters filed against legal research platform ROSS Intelligence. Thomson Reuters alleged that ROSS used a third-party bot to siphon copyrighted material from Westlaw in order to train its own AI systems. Yet, although copyrighted works such as law review articles could hypothetically work well as training data for AI-based systems, this may be an inefficient and unnecessary strategy. New sophisticated machine learning processes capable of unsupervised learning may make it possible to “train” legal search engines solely on non-copyrighted texts like briefs and statutory provisions. Technique and careful curation of data appear as equally essential factors in AI training that aren’t dependent on copyright. In other areas of the legal tech playground, copyright issues simply might not come up as often. Contracts tech development can rely on the many samples openly available in the market as a result of litigation, as well as greater cooperation from clients in exchange for promises of better results. Meanwhile, e-discovery providers have already adapted to training their AI applications on the publicly-available Enron data sets.
In August of 2019, a consumer launched a class action against Apple and Samsung for failure to provide warnings about radiation from their phones. Samsung was soon dismissed from the case, but the litigation against Apple continued. Although the iPhone radiation tests commissioned by plaintiffs showed levels that surpassed federal safety limits, the Federal Communications Commission tested both model and publically-available phone versions and found the radiation levels to be squarely within the mandated limits. Persuaded by the latter, a U.S. District Court dismissed the case, noting the FCC’s broad authority to enact uniform radiation emissions regulations, and looking to broader efficiency and uniformity.
The U.S. government is suing for the forfeiture of more than $1 billion in Bitcoins and other variants of cryptocurrency, linked to the illegal marketplace Silk Road that it shut down seven years ago. The 69,370 Bitcoins, seized by the Justice Department on November 3rd, represent the largest cryptocurrency bust in history. The announcement comes amid a resurgence in Bitcoin, which jumped almost to a three-year-high valuation of $15,300. The U.S. authorities claim to have seized the Bitcoins from an alleged hacker, identified as Individual X in the court documents, who successfully hacked the Silk Road sometime between May of 2012 and April 2013. Until its downfall in October of 2013, Silk Road was labeled by the U.S. government as “the most sophisticated and extensive criminal marketplace on the Internet” and has generated an equivalent of $1.2 billion in illicit sales in less than three years.
Snap Inc. was found not liable to a man who was severely injured by a driver allegedly distracted by a speed-recording feature on the company’s Snapchat app because it didn’t owe him a duty to change the product’s design, a split Georgia appeals court ruled. The defendant allegedly crashed into the plaintiffs’ car while trying to get the app’s speedometer filter to record a speed of 100 miles per hour. The plaintiffs alleged that Snapchat’s design contains an inherent incentive to engage in risky behavior. However, the court noted that the plaintiffs failed to point out any reward system in Snapchat predicated on misusing it while driving. Despite the alternative argument by the dissenting judge, the court stated that Georgia law does not impose a general duty on manufacturers to prevent people from committing torts while using the manufacturer’s product.
The U.S. Patent and Trademark Office is conducting a study on how often states infringe patents and trademarks without an adequate remedy under state law. Attempting to learn, among other things, how states and state entities usually respond to credible patent and trademark infringement claims, the agency will seek public comments as part of the study. Sovereign immunity remains a topic of interest in the intellectual property space, especially following the recent decision in Allen v. Cooper, where the Court “made clear that Congress does have authority in certain circumstances, to strip states of the immunity they enjoy” under the U.S. Constitution. The parallel investigation will explore a potential blueprint for the abrogation of sovereign immunity in certain copyright infringement claims.
A sports news site’s embedding of an Instagram post by professional tennis player Caroline Wozniacki was protected from a photographer’s copyright claims by the fair use doctrine, according to a Brooklyn federal court ruling. The court held that the site’s use of a social media post to report on that post transformative, adding that the news site argument did not reproduce the photograph to show Wozniacki as a tennis player, but only embedded the Instagram post “because ‘the fact that (Wozniacki) had disseminated’ that post ‘was the very thing the article was reporting on.’” Judge Ross limited this decision, stating that rather than giving publishers free rein to copy and paste images from Facebook or Instagram, it draws a line that balances photographers’ interest in protecting their copyright with reporters’ inters in covering social media events.
California voters have approved the California Privacy Rights Act (CPRA), which was listed on the ballot as Proposition 24. The CPRA imposes additional obligations on and considerations for businesses subject to the California Consumer Privacy Act, including providing consumers with the right to prevent businesses from sharing their personal information with third parties for behavioral advertising, establishing a new privacy protection agency in California dedicated to the implementation and enforcement of the CCPA and other California privacy laws, adding emails and passwords/security questions to the categories of personal information subject to the CCPA’s private right of action, and Recognizing “sensitive personal information” as a new category of personal information and requiring businesses to disclose how they collect. Businesses have until January of 2023 to comply with the new law. However, CPRA is much more complex than CCPA and will require a significant investment of time and money to convert existing systems to meet the new requirements.
Of the 1,369 legal tech patents filed with WIPO in 2019, 847 were filed by applicants in China. China’s dominance is likely to continue as its government encourages court innovation, and artificial intelligence maintains a significant role in legal tech software. Dina Blikshteyn of Hanes and Boone notes that “a lot of legal tech tends to relate to AI, and China is the biggest filer of AI in 2019 and 2020.” Most of the legal tech patents filed in China were for tools that enhance evidence collection and administrative tasks, aimed to counteract deficit in the number of lawyers for all citizens. Moreover, lower thresholds for patentability, for example, the ability to patent a business step, could extend China’s dominance in legal tech patent filings in the coming years.