Fake election views and AI-assisted prices expose two ways numbers can mislead
I have caught myself assuming a video with a huge view count must matter. That is exactly the shortcut an influence operation can exploit. Bloomberg reviewed 130 fabricated videos and doctored newspaper pages targeting US midterm candidates between September 10 and October 9. Researchers attribute the network to the Russia-linked Matryoshka operation; the reported sudden view spikes suggest automated amplification, not proof that hundreds of thousands of voters believed the claims. The fake clips impersonated celebrities and news broadcasts. Separately, the Associated Press examines growing anxiety about AI-assisted retail prices. These are not one coordinated campaign. Their common feature is that a number on a screen can feel objective while the process producing it remains hidden. A digital shelf label merely changes how a store displays a price. Dynamic pricing changes prices with conditions. Personalized or surveillance pricing uses information about a shopper to tailor offers; the FTC says vendors have offered such tools, but the prevalence and effect at particular stores remain uncertain. Walmart says its electronic labels do not price individual people. We should not turn a warning about possible practices into an accusation against every retailer. In elections, ask who created and amplified a clip before treating views as public opinion. At checkout, ask whether your own data shaped the price and whether you can see the basis for it. The next useful evidence is platform-level reach and engagement data for the reported bot network, and independent price audits that compare otherwise identical shoppers. Until then, the most honest conclusion is about opacity, not a measured change in votes or grocery bills.