New Video from @Computerphile Explores "Reputation Lag Attack" in Online Systems

The video from @Computerphile explores the concept of a "reputation lag attack" in online systems where reputation plays a crucial role. These systems include e-commerce platforms like Amazon or eBay, social networks, and even more obscure systems like the dark web or cryptocurrencies. Reputation is essential for maintaining trust between users, and any misconduct can lead to a deterioration of this reputation. "Reputation lag" refers to the delay between a bad action and the update of the user's reputation in the system.

For example, if a seller promises to deliver a product but does not do so immediately, they can delay bad reviews by providing excuses. This delay is natural and varies depending on the system. In a centralized system like Amazon, the reputation can be updated quickly after a negative review, while on social networks, the spread of a bad reputation may take longer. The "reputation lag attack" exploits this delay to maximize malicious actions before the reputation is affected. Attackers can use various strategies to extend this delay or to perform as many malicious actions as possible in a short period.

For example, an attacker could advertise a very attractive product at a low price to attract many buyers before their bad reputation is discovered. The video also mentions other types of reputation-related attacks, such as the "bad mouthing attack," where negative reviews are left on competitors, or the "whitewashing attack," where a user creates a new account after tarnishing their reputation. Another well-known attack is the "Sybil attack," where users create multiple accounts to support each other with fake positive reviews.

The impact of the network structure on these attacks is also discussed. For example, in a social network, influential nodes with many connections can spread a bad reputation faster than peripheral nodes. However, research shows that this difference is often more subtle than expected. Finally, the video addresses the example of Honey, a browser plugin that tarnished its reputation by reducing the earnings of its promoters. This example illustrates how a bad reputation can quickly destroy user trust, even if the initial intention was not malicious. In conclusion, understanding and managing "reputation lag" is crucial for maintaining trust and security in online systems. Companies must be aware of these dynamics to protect their users and their own reputation.