OpenCV can detect faces and/or upright people and has examples of doing both. It can be used from C/C++/python and is open source & cross platform.
While it is not a ready made solution it can be used to construct such solutions as can be seen with the examples of pedestrian detection & face detection presumably you would need to do both to build a "people" detector.
OpenCV + Python & Keras can also be used to produce a "deep learning" classifier using TensorFlow or Theano where the computer teaches itself how to categorise images and there are some pre-trained models available which may be able to classify your images for you, see here.
The above resources should enable you to create a script that will do what you need.
Disclamer: While most of the links above point to a single online resource that has a commercial element, training, that is that I am a fan because I find it a great resource not because I have any financial interest in the site or any of the services offered.