Vice President Kamala Harris announced a swath of new initiatives at the United Kingdom’s Global Summit on AI Safety Wednesday, including the establishment of the United States AI Safety Institute. One of the initiatives outlined by the White House is to combat fraudulent robocalls using artificial intelligence.
The White House will host virtual hackathons to fight fraudulent spam callers, particularly those using novel AI-generated voice models. VP Harris invites technology experts around the country to build artificial intelligence-based solutions to fend off the scams that particularly target the elderly. This initiative, among others announced by VP Harris Wednesday, expands on Joe Biden’s executive order calling on Congress to regulate artificial intelligence.
The summit hosted by UK Prime Minster Rishi Sunak started Wednesday, featuring government officials from around the globe focused on preventing the worst outcomes of artificial intelligence. The summit lacks global leaders, as Canada’s Justin Trudeau, France’s Emmanual Macron, and Joe Biden have all sent representatives on their behalf. Harris and Tesla CEO Elon Musk, among others, are at the UK summit to discuss the risks of advanced large language models that have caught regulators worldwide on their heels.
The Biden Administration is also calling on all nations to develop international watermarking standards that help the public identify authentic government digital content versus AI-generated content. AI deep fakes have plagued internet users for years, and social media platforms have been struggling to combat them.
A political declaration around military AI capabilities has been signed by 30 nations, including the UK, Sweden, and the United States according to VP Harris. The declaration sets norms for responsible development and deployment of AI technology in military settings.
Sunak’s AI summit is expected to focus on long-term, worst-case scenarios in artificial intelligence, rather than issues facing large language models today. Several large language models are facing copyright infringement lawsuits, from OpenAI to Google, however, the topic does not appear to be on the docket this week. Instead, conversations around cyberattacks, disinformation campaigns, and how advanced AI could undermine human control will take place at the summit.
