Synthetics, sometimes called 'androids', AIs, or 'synths', is a loose categorization of sentient machines with autonomous protocols, learning heuristics, and a generally bipedal/humanoid appearance. All modern synthetic life-forms descend from the original design by Nikolai Tesla, and his original android Atomic Robo.
True sentient artificial intelligence, defined by an ability to learn and adapt, is extremely difficult to reproduce. Tesla's research into Zero-Point Energy was so forward-thinking that elements of it have still not been reproduced reliably. Atomic Robo was Tesla's sole success. One of Tesla's junior engineers, Phineas Horton, stole a prototype Z-point construct and used it to build a synthetic life form dubbed The Human Torch. The Torch was an active component of the postwar Justice Defenders until self-terminating in 1955 due to increasing instability in his Z-point construct.
Tony Stark successfully produced several fully digital AI constructs for Stark Industries, first JARVIS, then later, HOMER. Rather than being stored in a mobile 'brain', they are largely distributed intelligences that require vast infrastructure and quantum storage to achieve requisite cross-sectional neural density.
In 2014, Dr. Hank Pym was approached by the Department of Defense to create a synthetic life form for the White House. This was code-named the Ultron project. The project's results were so catastrophic that Pym single-handedly destroyed all project files, resources, and records, setting back synthetic research by decades. The construct known as Vision and later, Vivian Vision, were two of the successes of that program.
'Proto-Sapient AIs', also called 'Virtual Intelligences', have become significantly more common as computer design evolves. These designs lack the ability to modify personality constraints or edit their learning heuristics, and thus not truly 'sentient'. These VIs are most commonly seen in the use of Life-Model Decoys (LMDs) as well as advanced technology such as aviation, building management, and network administration.