OpenAI prepares IPO, lobbying drive

OpenAI was preparing to file for an initial public offering within weeks as of May 20, according to the New York Times, while the company was also pressing a state-level lobbying campaign to shape AI rules in the absence of federal legislation. Politico reported the same day that OpenAI’s policy operation was working through state capitals, especially Democratic-led ones, to build support for a national framework. The push came as a U.K. study found ChatGPT and other chatbots gave false election information before Scotland’s May 7 parliamentary vote, and as President Donald Trump was expected to sign a U.S. executive order on AI and cybersecurity as soon as May 21. ### Why is OpenAI trying to do an IPO now? The New York Times reported on May 20 that OpenAI was preparing to file to go public in coming weeks, in what would be one of Silicon Valley’s most closely watched offerings. Bloomberg and CNBC separately reported that the company was working with Goldman Sachs and Morgan Stanley on a confidential filing that could come as soon as Friday, though the timing remained uncertain. (nytimes.com) CNBC reported that an OpenAI representative said the company’s focus “remains on execution.” Bloomberg said the company was targeting a public debut sometime in the fall, according to a person familiar with the plan. ### What is the lobbying campaign trying to achieve? Politico reported on May 20 that OpenAI’s top lobbyist, Chris Lehane, was pushing what he called “reverse federalism,” a strategy to advance AI policy through states while Congress remained deadlocked. (nytimes.com) The outlet said OpenAI was increasingly spending time in statehouses to support laws that could align with a national framework. (cnbc.com) Washington’s slower pace has been a central part of that argument. Politico said the industry’s push for tech-friendly federal legislation had foundered in Congress, prompting OpenAI to pursue state-level action instead. ### What happened in Scotland? The Guardian reported on May 20 that a Demos study found ChatGPT and other AI chatbots produced fabricated scandals, invented candidates, gave the wrong election date and falsely said Scottish voters needed identification at polling stations. (politico.com) The study was published after Scotland’s parliamentary election on May 7. The Electoral Commission responded by calling for new controls, according to the Guardian. Other reports summarizing the Demos findings said 34% of chatbot responses contained errors. ### What is Trump expected to do? Reuters reported on May 20 that Trump was expected to sign an executive order on AI and cybersecurity as soon as Thursday, May 21. (article.wn.com) The order would create a voluntary framework for AI developers to engage with the U.S. government about releasing covered models, according to two sources familiar with the matter. Reuters also reported that the White House was working to bring AI company chief executives to a signing ceremony. Earlier reporting from Bloomberg and CNBC had described administration discussions about a model-testing or vetting system, but Reuters said the expected order would stop at a voluntary framework. (usnews.com) ### How do these pieces fit together? May 20’s reporting showed OpenAI moving on two tracks at once: preparing for public markets and trying to shape the rules of the market it will enter. Politico’s account of state lobbying and the Scottish election study both pointed to the same pressure point — regulators are being asked to respond before Congress settles on a comprehensive law. (usnews.com) The next concrete steps were close at hand. CNBC said OpenAI could confidentially file as soon as Friday, while Reuters said Trump’s AI order could be signed on May 21 if White House planning held. (cnbc.com) (nytimes.com)