Bubbles are omnipresent in lab experiments with asset markets. But these experiments were (mostly) conducted in environments with only human traders. Today markets are substantially determined by algorithmic traders. Here we use a laboratory experiment to measure human trading behaviour changes if these humans expect algorithmic traders. To disentangle the direct effect algorithmic traders have we use a clean design where we can manipulate only the expectations of human traders. We find clearly smaller bubbles if human traders expect algorithmic traders to be present.