Theft and deceit of virtual assistant
The development of neural networks has allowed mankind to reconsider computing. Neuromorphic computers (neuromorphic computing) that imitate the work of the human nervous system have emerged. They store and process data using the same elements, and the capabilities of such computer are determined by the dataset “fed” to it.
The emergence of these computers has brought about unprecedented opportunities: now everyone can get one’s own digital entity - a personal assistant that owns all the data about a person collected since his very birth. This is a kind of a second digital "self".
Such an entity can make decisions on behalf of the owner, because it acts the same way as its owner, so that people have started to delegate many routine issues to their digital avatars.
An interesting feature of neuromorphic computer consists in impossibility to run malware. It can be cheated in two ways only:
1) Steal the entire system (copy), and thus learn all about the person (its owner). Including how he/she will behave in a given situation.
2) Give a wrong dataset to the avatar
More details about the second scenario:
Neural network underlying the digital essence, draws upon the data set while making decisions, which is, in fact, the past of the person. If this data set is substituted, the neural network will make decision that isn’t typical of the person. Thus, a virtual entity that is, for example, often used for business negotiations, can make a decision that a person would never make. But as it is considered a full-fledged representative of the person - in fact, to some extent, it can be regarded as this person – its decision has the same power as its owner’s.