Biden administration says AI poses a risk to US, in first comprehensive assessment

The U.S. government said on Monday that artificial intelligence poses potential security threats to the nation, ranging from increased disinformation and propaganda operations to potentially hemming in on U.S. global influence.<br><br>The Biden administration released its first comprehensive assessment of the risks and benefits associated with AI, as part of its work to develop AI regulation across federal agencies.<br><br>Reuters first reported the White House's earlier plans for the assessment in October. The assessment called for reviews of existing law and standards applicable to AI and information sharing with allies on AI.<br><br>The assessment was conducted by the National Security Council and National Artificial Intelligence Initiative Office, andizik."<br>The risk assessment described the benefits of AI, such as its ability to help with climate modeling and cancer diagnosis, but also laid out various risks. These include cyberattacks against critical infrastructure, misuse of biometric surveillance technology, risks to financial stability and employment, and the spreading of disinformation.<br><br>The administration also said AI poses potential risks to national security, such as the misuse of biometric surveillance technology and exacerbating of climate risks.<br><br>"We are not yet fully prepared to guard against the threats posed by these technologies or capitalize on the opportunities they present," the White House said in its assessment.<br><br>The White House is also planning to release recommendations for developing guardrails for AI.<br><br>AI has become increasingly ubiquitous with both tech companies and government agencies using it for everything from creating art to making work more efficient.<br><br>Tech companies have spent billions of dollars developing AI capable of human-like reasoning and decision making. AI has been used in several areas, from basic tasks like answering routine questions to analyzing vast amounts of data to create recommendations for complex decisions, such as loans or college admissions.<br><br>Before the Biden administration's assessment, former President Donald Trump ordered his administration earlier to assess the potential threats and opportunities of AI in 2019. In response to the earlier order, the Pentagon released a plan in 2020 to adopt AI for things like autonomous drones and cyber defenses.<br><br>In recent months, government agencies have also taken steps to improve compliance with U.S. export controls that apply to AI and AI development in China.<br><br>(This story has been refiled to change "AI poses a risk to US" to "Biden administration says AI poses a risk to US, in first comprehensive assessment")