2.4.3. Fraud

Fraud represents the potential risks of personal energy data being modified without authority, either to increase/decrease energy consumption or attribute the energy consumption to another

house [32]. This risk belongs to category ii; the AMI enables more opportunities for adversaries to implement fraud than conventional meters since the vulnerabilities of the real-time communication network would be abused.

#### 2.4.4. Non-Grid Commercial Uses of Data

This privacy risk falls into category ii. The smart meter data may be used by TP to make a profit from the data; activities include advertising and insurance that are not welcomed by consumers [32]. Companies would sell their products to residents according to the personal preference information revealed by the energy data. Even sensitive information, such as employment information [30], can be inferred from energy data with machine learning algorithms. This information can be used by adversaries to estimate the income of the target family.

#### **3. Related Work for Privacy Intrusion Protection**

The state-of-the-art methods dealing with the above smart meter privacy issues can be divided into two categories: user demand shaping and data manipulation. Both these techniques try to reduce the privacy loss by decreasing the probability to infer individual appliance signatures from the overall power data [34].
