In January, a Knight-Ridder reporter blanketed the nation with his byline under the headline "AIDS Cases May Be on the Rise After Seven Years of Decline." Poring over preliminary data from the CDC, Seth Borenstein spied an 8 percent increase from 38,864 in 2000 to 42,008 in 2001. Then he called around for predictably dark comments about unsafe sex. What he didn't do was check his facts with the CDC. The feds fired off a letter debunking the hype, dismissing the numbers as "provisional weekly data, not a reliable indicator of trends." Adding insult to injury, the stats are a running count of HIV transmissions, not AIDS diagnoses. Nor had they been adjusted for duplicates and other errors. But the CDC, for its part, is throwing more attitude than light: All officials will share with POZ is that HIV rates remain at a steady 40,000 -- as they have for seven years.