AI News Roundup – Canadian news organizations sue OpenAI, Trump considers appointing “AI czar,” fighting scammers with AI, and more

To help you stay on top of the latest news, our AI practice group has compiled a roundup of the developments we are following.

• • A group of Canada’s major news organizations is suing OpenAI, alleging illegal use of the group’s content in training the company’s AI models, according to The Washington Post. The coalition, which includes the newspapers The Globe and Mail and the Toronto Star as well as the Canadian Broadcasting Corporation filed a lawsuit in Ontario Superior Court this past week, claiming that OpenAI illegally scraped their copyrighted content without permission or compensation. The coalition is seeking 20,000 Canadian dollars, or $14,700, per article they claim was illegally scraped and used to train OpenAI’s ChatGPT, as well as a share of OpenAI’s profits derived from the alleged illegal use of their content. This latest legal action mirrors similar lawsuits in the U.S., including a suit brought against the company by The New York Times, and reflects a growing international dispute over AI companies’ practices of using vast amounts of online content to train their algorithms.

• • U.S. President-elect Donald Trump is considering appointing an “AI czar” to oversee his incoming administration’s AI regulation and use efforts, according to Axios. The potential White House role would focus on coordinating federal policy and governmental AI technology use, with significant input from Elon Musk and Vivek Ramaswamy through their newly formed Department of Government Efficiency (DOGE). The AI czar would be responsible for leveraging both public and private resources to maintain U.S. technological leadership, working with agency chief AI officers to identify opportunities in government and promoting private investment in AI technologies. While Musk will not directly hold the position, he is expected to play an influential behind-the-scenes role in forming the second Trump administration’s tech policy. Such an AI czar appointment would not require Senate confirmation, allowing for faster implementation of the administration’s AI-related goals, and might potentially be combined with a cryptocurrency leadership role under a broader emerging technologies czar.

• • The British telephone company O2 has unveiled a new AI-generated granny designed to waste the time of scammers who often prey on the elderly through phone calls, according to The New York Times. Named Daisy Harris, this AI character is programmed to engage phone scammers in meandering, time-consuming conversations, drawing inspiration from real grandmothers’ conversational styles and technological naivety. Developed by O2 in collaboration with an advertising agency and a professional scammer, Daisy is meant to appear as an ideal scammer target, with a backstory involving birds, a cat named Fluffy, and a knitting hobby. While experts acknowledge that Daisy represents only a small disruption to the massive global phone scam industry—which sees tens of millions of scam calls daily and over $1 trillion stolen annually—the AI character can potentially divert scammers’ attention by keeping them on lengthy, unproductive calls. However, criminology experts still recommend that individuals simply hang up and report fraudulent calls rather than attempting to engage with scammers.

• • Anthropic has introduced a new open-source tool to connect AI models with data sources, according to The Verge. The Model Context Protocol (MCP) is designed to provide a universal connection method for AI systems to access various datasets, eliminating the need for developers to create custom code for each data source. Alex Albert, Anthropic’s head of Claude relations, explained that the tool allows developers to integrate their AI systems with a standard protocol for sharing resources, tools and prompts, making it easier to connect AI agents with multiple data sources. In a blog post, Anthropic explained that the MCP also provides AI systems with the ability to “maintain context as they move between different tools and datasets, replacing today’s fragmented integrations with a more sustainable architecture.” Companies like Replit, Codeium, and Souregraph have already begun implementing the MCP, and the company is seeking feedback from developers who implement the new tool into their AI applications.

• • London’s Heathrow airport is testing a new AI tool designed to aid air traffic controllers in managing the flow of aircraft to and from Europe’s busiest airport, according to Simple Flying. Named “Amy,” the AI system has already been tested on over 40,000 flights, integrating radar and video data to monitor aircraft across the airfield and providing real-time visualization of aircraft positions. This innovation comes at a critical time when Europe is experiencing significant air traffic controller shortages, with 2023 being the worst year in two decades for air traffic control performance. The AI is specifically designed to work alongside human controllers, not replace them, and can track flights from touchdown to runway exit, helping to improve safety and efficiency at an airport that handles approximately 1,245 flights per day. The testing of Amy is part of a broader effort to address ongoing challenges in air traffic management, with airlines and industry groups calling for urgent reforms to address capacity and service quality issues.