OpenAI helps spammers plaster 80,000 sites with messages that bypassed filters
1 min read
Summary
Hackers have used open AI to develop a spambot that was able to generate messages that bypassed spam filters and were tailored to the recipient.
The bot, called AkiraBot, was used to send promotional messages for shady search optimisation services to more than 80,000 websites.
The malicious activity went on for four months before the OpenAI subscription held by the hackers was revoked.
The case shows both the benefits and dangers of AI, as the technology can be used for both good and malicious purposes.
AI can synthesise large amounts of data and create content at scale, but that can also mean the speed and breadth of malicious behaviour is increased.
The case also shows the reactive nature of many companies in responding to breaches and threats, as the four-month duration of the spamming could have caused damage to many websites during that time.
