US Patent and Trademark Office to hold listening session on AI's contribution to inventions, including ‘AI inventorship issues’

• The US Patent and Trademark Office (PTO) is holding a listening session on the contribution of artificial intelligence to new inventions on April 25th.

• The PTO will address the current state of AI technologies and inventorship issues, including whether AI should qualify as an inventor and whether the PTO should expand its current guidance.

• The PTO is taking a closer look at its current policies to see if they require updating to incentivize AI innovation, as AI systems continue to rapidly improve.

• The PTO previously created the AI/ET Partnership last June to address emerging technologies issues, following a court battle over the PTO's refusal to list computer scientist Stephen Thaler's AI system DABUS as an inventor on a patent in 2020.

• The PTO issued a request for comments on AI inventorship and AI-assisted works in February and the upcoming listening session will focus on questions put forth in that request for comments.

Samsung data leak highlights mounting privacy concerns surrounding ChatGPT use by employees

• Samsung employees reportedly leaked sensitive data to ChatGPT on three separate occasions.

• The information shared with the chatbot allegedly included the source code of software used to measure semiconductor equipment, and an excerpt from a corporate meeting.

• The leak occurred 20 days after Samsung lifted its ban on ChatGPT, which had been put in place to prevent the leaking of confidential data.

• Privacy concerns over ChatGPT’s security have been ramping up since OpenAI revealed that a flaw in its bot exposed parts of conversations users had with it, as well as their payment details in some cases.

• ChatGPT's data privacy concerns extend beyond Samsung. According to recent research by Cyberhaven, 3.1% of its customers who interacted with the AI chatbot submitted confidential company information to the system. This suggests that even organizations with 100,000 employees could be unknowingly sharing confidential data with OpenAI numerous times per week.

• Other major companies such as Amazon and Walmart have issued warnings to their employees about sharing sensitive information with the AI chatbot. Verizon and J.P. Morgan Chase have completely blocked the tool for their employees.

EU-US Data Privacy Framework and AI regulations remain works in progress

• The IAPP Global Privacy Summit 2023 in Washington, D.C. brought together data privacy stakeholders from academia, technology companies, the White House, and the European Union Parliament.

• Panelists discussed ongoing transatlantic negotiations for the successor to the Privacy Shield Agreement and the legal frameworks needed to account for advancements in technology.

• The European Data Protection Board (EDPB) commended the efforts of negotiators but flagged concerns about individual rights not covered by the new data privacy framework, such as the right to data portability.

• The U.S. Department of Justice has established a Data Protection Review Court (DPRC), but the EU recommends withholding the issuance of the final adequacy decision until the U.S. intelligence agencies effectively implement the policies and procedures needed to comply with the executive order.

• Panelists from both the U.S. and EU discussed the challenges of governing artificial intelligence and the different approaches being taken.

• The EU’s proposed AI Act is still making its way through the legislative process, while the U.S. has taken a less regulatory approach, focusing on existing legal frameworks and the recently released Blueprint for an AI Bill of Rights.

Legal tech companies Fastcase and vLex merge to create ‘the biggest legal data corpus ever assembled’ for AI products

• Fastcase and vLex are merging to create vLex Group, in a deal that will speed up the creation of AI tools for lawyers.

• The merger creates the biggest legal data corpus ever assembled, with over one billion legal documents from more than 100 countries, including judicial opinions, statutes, regulations, briefs, pleadings, and legal news articles.

• The companies are betting that their combined document library will be a rich training data set for legal AI products, as large language models like OpenAI's GPT-4 and Google's Bard quickly increase their power and reach.

• The combined company will license the dataset to other entities for use in creating their own large language models, and will build its own products trained on the legal data.

• Read the press release here