A few ponder the possibility of Artificial Intelligence running out of control, causing havok. Possible? Maybe. I think it unlikely. At least until software can be consistently designed, written, and revised with zero flaws.
The number of bugs in software overall remains fairly high. Apps, web browsers, operating systems, and various programs update monthly/yearly as a consequence of discovered flaws or limitations.
The concept of Bounded Rationality is partly at work.
Either advanced AI will be helped by having flaws the way people do or advanced AI requires perfection. On the other hand, perhaps perfect, flawless software is too limiting a vessel for any AI to emerge purpose contrary to people. Current technology (hardware, metals, electricity) seems too primitive and confining to support the existence of truly advanced AI. I think it would short-circuit before it could proceed too far.
Biology on the other hand is definitely more advanced than electronics in terms of the real world. Real life viruses like Zika has already had a major impact in the death of millions of bees. That will impact the food hierarchy and many other things.
People are capable of doing substantial harm to others. Gene-editing has arrived with Chinese scientists to pioneer gene-editing trial on humans. A technique called CRISPR makes this possible. CRISPR enhances genome engineering. Adapting your genes to produce effects on the body. It is a huge advance because it can potentially reduce diseases and cancers. However, it may make it possible to see the emergence of new humans. People resembling Friedrich Wilhelm Nietzsche’s Übermensch in various ways. Being smarter, faster, more charismatic, and richer than the rest of us, they’ll have a far more substantial impact than AI.