A hazard curve tells you what the probability is of any particular strength of ground shaking. It doesn't tell you which value you should choose to design your building against. Do you want to be 95% safe, 99% safe, or 99.9% safe? These are really economic or political decisions, not seismological ones.
Also, one has to bear in mind that low probability events do happen. The Maharashtra earthquake of 1993 is a good case in point. If a seismologist had been assessing the hazard in this part of India in 1992, he would have concluded that the probability of a damaging earthquake was extremely low. And he would have been right. Unfortunately, that very small probability came up next year.
Consider a game in which you have one chance to roll six dice. If you roll six sixes, you have to pay $100,000. Any other result and you win $1,000. A quick calculation shows that the game is heavily stacked in your favour. But that doesn't mean you can't lose. Seismic hazard assessment helps the engineer stack the odds in his favour. To be 100% safe you would have to design every building against improbably large earthquakes occurring unexpectedly close to your site almost irrespective of whether you were in a high activity zone or not. And you would still have to worry about that 50-ton meteorite scoring a direct hit...