The downside of Arduino is that it discourages learning how things really work. It teaches you sorta-kinda how things work while shielding you from the real challenges and at the same time making the user feel like an engineering genius.
That's an idea that's often stated, but are there any studies to back that up with evidence? Huge numbers of people are never going to care who things really work, as they just lack curiosity. Does Arduino really discourage the naturally curious from probing deeper? That seems like the only way it might have a negative effect.
+1 for coppice's comments
I think it's like saying real basketball players play on 10' goals. Kids (including those who grow up to play in the NBA) start on 8' goals (and lower), and they use a 28.5cm basketball until about 7th grade before moving to a 29.5cm ball.
Additionally, the world is evolving. Technologies and tools that are available to kids today weren't available to people who are adults now when those adults were kids - so some adults are learning from an elementary level.
If an Arduino helps someone become interested in electricity or hardware or software, what's the problem? It doesn't mean they will never learn anything else or not move on to bigger challenges.
Some kids who learned basketball on an 8' goal never got around to playing on a 10' goal but many move up to the more challenging height as they gain experience. I have never heard of a kid who learned to shoot on an 8' goal who then said "that's too easy so I'm going to quit basketball and never shoot on a 10' goal" and I never heard of a kid or parent or a coach who said "because we started with a smaller ball and a lower goal our curiosity about basketball techniques, understanding of basketball, enthusiasm for learning basketball, and basketball skills were ruined before we could get a chance to experience the bigger ball and higher goal." Just doesn't happen.