Consider a purely hypothetical example: over one year, an event occurs in 2 out of 100 people in one group and 3 out of 100 in another. The relative increase is 50%, because 3 is one and a half times 2. The absolute difference, however, is one percentage point: one additional person per 100 over the same year.
Both numbers are correct. Relative risk describes the ratio between the two risks; the absolute difference shows how much the frequency of the event changes.
When a headline presents a percentage, also look for the starting risk, the time period and the population it describes. These numbers help you understand the size of a finding; they cannot, on their own, tell you how reliable the study is.