Google fights AI manipulation

Google has spent the past week tightening rules around its AI search products after a BBC investigation found simple ways to influence chatbot-style answers with false or self-serving material. On May 15, Google updated its search spam policy to say spam includes attempts to manipulate “generative AI responses in Google Search,” extending rules that already covered efforts to game traditional rankings. (developers.google.com) The policy change landed as a separate backlash played out on college campuses. (vuink.com) NPR affiliates and other outlets reported that commencement audiences booed speakers who praised artificial intelligence, including at ceremonies where graduates said AI talk sounded detached from fears about jobs and credibility. (blog.google) Together, those two developments describe the same problem from different angles. Google is trying to show its AI answers can be defended against manipulation, while students are showing that public patience with boosterish AI language is thin. (developers.google.com) ### What, exactly, did Google change? Google’s Search Central documentation now defines spam as techniques used to deceive users or manipulate search systems into featuring content prominently, including attempts to manipulate generative AI responses in Google Search. The updated language appears in the company’s public spam policies and applies to AI-generated features such as AI Overviews and AI Mode, according to Search Engine Land’s review of the change. (kgou.org) Search Engine Land reported on May 15 that Google had clarified the rules after a growing market for tactics aimed at winning placement inside AI answers rather than only in blue-link results. Google’s documentation says sites that violate spam policies can become ineligible to appear prominently in Search. (developers.google.com) ### How was Google’s AI being manipulated? The BBC reported on May 19 that its investigation found a simple route to getting chatbots to repeat misinformation. The available BBC synopsis says researchers and reporters showed how AI systems could be pushed into surfacing false claims to the public, prompting criticism and follow-up fixes by Google and others. (developers.google.com) Google has not, in the sources reviewed here, published a point-by-point response to the BBC’s specific test cases. But its public policy update and product posts show the company moving on two tracks at once: penalizing manipulation attempts and emphasizing links to outside sources inside AI answers. (searchengineland.com) ### Why are graduates booing AI at commencements? University of Central Florida graduates booed after commencement speaker Gloria Caulfield said, “We’re using a new AI system as our reader,” according to KGOU’s account of the NPR reporting. Slate separately reported that former Google chief executive Eric Schmidt drew boos at the University of Arizona after referring to Time magazine’s “Architects of AI” issue. (vuink.com) The Associated Press version of the story, carried by Fortune on May 20, described the reaction as part of an “AI summer of 2026” in which young graduates are weighing new tools against worries about entry-level work. KGOU said commencement speakers who mention sweeping AI-driven change are encountering resistance from the Class of 2026. (developers.google.com) ### What does this mean for schools using classroom AI tools? Google’s own product posts say AI Overviews and AI Mode are experimental and that the company is still improving how those features show links and connect users with “authentic voices.” That matters for schools because classroom adoption often depends on claims that AI outputs are reliable, explainable and easy to verify. (kgou.org) NPR’s broader reporting on higher education has also shown colleges struggling to decide how AI belongs in teaching. In that context, the combination of manipulation concerns and visible student hostility suggests schools should expect questions about where AI answers come from, how they are checked and why a tool is necessary at all. (fortune.com) That final point is an inference drawn from the policy change and the commencement backlash, not a statement made by Google or NPR. ### What happens next? Google’s published rules are already in effect, and the company says its AI search features will continue to evolve as it improves source links and discovery tools across AI Overviews and AI Mode. (blog.google) The next test will be whether those safeguards reduce abuse while schools and universities decide how much trust to place in AI systems that Google still labels experimental. (developers.google.com) (one.npr.org)