Philadelphia's embattled sheriff, Rochelle Bilal, is facing criticism after it was revealed that her campaign team had posted a series of positive “news” stories on their website that were generated by ChatGPT, a generative AI chatbot. The campaign team has since removed over 30 of these stories, following a report by the Philadelphia Inquirer that local news outlets could not find any evidence of these stories in their archives.

Experts have expressed concern that this kind of misinformation can undermine public trust and pose a threat to democracy. Bilal's campaign has claimed that the stories were based on real events, but it has been revealed that an outside consultant used the AI service to generate these articles based on talking points provided by the campaign.

Large language models like OpenAI's ChatGPT are designed to predict the most plausible next word in a sentence, allowing them to quickly complete challenging prompts. However, this also means that they can make frequent errors, known as "hallucinations." While many Americans have started using these tools to write documents more efficiently, the lack of accuracy and fact-checking can lead to problems.

This incident is not the first time AI tools have caused issues. Last year, two lawyers had to apologize to a judge in Manhattan federal court after using ChatGPT to search for legal precedents and failing to notice that the system had made some up. Mike Nellis, the founder of the AI campaign tool Quiller, has criticized the consultant's use of AI as "completely irresponsible" and "unethical."

OpenAI, the company behind ChatGPT, is responsible for enforcing its policies, which prohibit the sharing of output from its products to deceive or scam people. They also do not allow the use of their systems for political campaigning or lobbying. However, there is no evidence to suggest that such misuse occurred in this case. OpenAI has not yet responded to requests for comment.

The incident has highlighted the need for regulation of AI tools in politics as the technology continues to advance. Bipartisan discussions in Congress have emphasized the importance of such legislation, but no federal law has been passed yet.

The list of news stories on Bilal's campaign website, titled "Record of Accomplishments," ended with a disclaimer stating that the information provided makes no representations or warranties about its accuracy. This disclaimer, which the Inquirer described as new, has raised concerns among some, including a fired whistleblower from Bilal's office, who fear that misinformation could confuse voters and further erode trust in democratic institutions.

Brett Mandel, who briefly served as Bilal's finance chief in 2020, expressed his concerns about the situation, stating that there has been a disregard for truth and a loss of trust in both local and national levels. Mandel has filed one of several whistleblower suits against the office, alleging that he was fired for raising concerns about office finances.

Bilal has faced criticism throughout her tenure as sheriff, including allegations of excessive office spending, issues with campaign finance reports, and the reported loss of hundreds of weapons. The list of news stories on the campaign website attributed four stories to the Inquirer and others to local broadcast stations WHYY, WCAU, and KYW. However, the Inquirer has confirmed that none of these stories are in their archives.

This incident serves as a reminder of the importance of accuracy and responsible use of AI tools in political campaigns. It also highlights the need for regulations to ensure the proper use of AI in politics as the technology becomes more prevalent.