Google's AI is being manipulated. The search giant is quietly fighting back (www.bbc.com)

A recent BBC investigation uncovered alarming vulnerabilities in AI chatbots, including Google’s AI and ChatGPT, that allow users to manipulate these systems into spreading misinformation. The findings revealed that by publishing a fabricated but persuasive online article, the investigator was able to have the AI tools present false claims as credible information, impacting serious topics like health and finance. This manipulation poses a significant risk, as billions depend on AI-generated assessments, making them susceptible to biased or inaccurate content that could influence critical decisions. In response to these revelations, Google updated its anti-spam policies, signaling a renewed commitment to combat the exploitation of its AI systems. Although the company claims the changes are clarifications rather than a shift in approach, experts suggest that Google and its competitors are actively experimenting with new methods to filter out misleading content. These adjustments include adding confidence labels to AI responses and potentially removing sources suspected of self-promotion. However, skepticism remains among experts, who warn that as defenses evolve, so too do the tactics employed by those seeking to manipulate AI outputs, highlighting the need for continued vigilance in assessing AI-generated information.