"Contrary to conventional thinking, the risk of diabetes is higher with 'light' beverages compared with 'regular' sweetened drinks," the National Institute of Health and Medical Research (Inserm) said.
The evidence comes from a wide-scale, long-term study, it said in a press release.
More than 66,000 French women volunteers were quizzed about their dietary habits and their health was then monitored over 14 years. The women were middle-aged or older when they joined the study.
Sugar-sweetened sodas have previously been linked with an increased risk of diabetes, but less is known about their artificially-sweetened counterparts.
Researchers led by Francoise Clavel-Chapelon and Guy Fagherazzi dug into the data mine to look at the prevalence of diabetes among women who drank either type of soda, and those who drank only unsweetened fruit juice.
Compared with juice-drinkers, women who drank either of the sodas had a higher incidence of diabetes.
The increased risk was about a third for those who drank up to 359 millilitres per week of either type of soda, and more than double among those who drank up to 603 ml per week.
Drinkers of light sodas had an even higher risk of diabetes compared to those who drank regular ones: 15 per cent higher for consumption of 500 ml per week, and 59 per cent higher for consumption of 1.5 litres per week, Inserm said.
There was no increase in diabetes among women who drank only 100-per cent fruit juice, compared with non-consumers.
The study noted that women who drank "light" sodas tended to drink more of it -- 2.8 glasses a week on average compared to 1.6 glasses among women on "regular" sodas.
The findings are published in the latest issue of the American Journal of Clinical Nutrition.
Its authors admitted the study had limitations.
"Information on beverage consumption was not updated during the follow-up, and dietary habits may have changed over time," the paper said.
"We cannot rule out that factors other than ASB (artificially sweetened beverages)... are responsible for the association with diabetes."
The authors urge further trials to prove a causal link.
The study covered women born between 1925 and 1950, who have been monitored since 1990.
The paper noted previous research which says that aspartame -- the most frequently-used artificial sweetener -has a similar effect on blood glucose and insulin levels as the sucrose used in regular sweeteners.