Concerns about the potential criminal abuse of AI chatbots have reached a new height, with Europol warning that ChatGPT is already being used to commit crimes. The European Union’s law enforcement agency on Monday published a report detailing the potential ways in which AI language models can be used for cybercrime, fraud, and terrorism.
OpenAI-owned ChatGPT went viral around the end of 2022, quickly becoming a hit among internet users. From casual purposes like generating jokes to professional uses like programming code, ChatGPT has found widespread application. However, the threat of potential misuse has always loomed over the popular chatbot.
Europol’s report on the criminal use of ChatGPT
Europol’s report is a confirmation of the fears surrounding the potential criminal use of OpenAI’s AI language model. The report claims that people have already begun using ChatGPT for nefarious purposes.
Criminals are typically quick to exploit new technologies and were fast seen coming up with concrete criminal exploitations, providing the first practical examples mere weeks after the public release of ChatGPT.Europol Report
The law enforcement agency highlighted how criminals could leverage ChatGPT to learn how to commit crimes that are outside their area of expertise. Even those who know nothing about a particular crime can ask the AI chatbot for a step-by-step guide.
Potential crime areas on which the language model might guide criminals include home invasion, terrorism, child sexual abuse, cybercrime, and more.
ChatGPT is admittedly better than other language models at blocking out potentially harmful input requests. However, loopholes still exist, and users have found ways to go around the safeguards implemented by the content filter system. The fact that some users managed to make ChatGPT provide them with instructions on how to make crack cocaine or pipe bombs.
Intensifying an existing problem
The report also admitted that all the information that users can obtain from ChatGPT already exists on the internet and is publicly available. However, the advanced AI chatbot makes them far easier to access. It significantly cuts down the research time needed to learn about a specific crime and provides criminals with detailed step-by-step guides.
Propaganda and misinformation generated using the language model might be particularly used to facilitate terrorism.
Besides learning the ropes to new crimes, criminals may also use ChatGPT to perform malicious tasks like impersonating others, producing propaganda, and phishing.
ChatGPT’s ability to generate codes isn’t to be ignored either. While it makes life easier for programmers, it also means cybercriminals with minimal technical knowledge can exploit it to create malicious codes.
The LLM would be even more dangerous in the hands of more advanced users, allowing them to refine their techniques. To make things worse, they might even use ChatGPT to automate criminal activities.
ChatGPT’s sophistication is a cause for concern too
ChatGPT and its underlying language model have also been incorporated into various services and applications, including search engines. The recent rollout of ChatGPT plugins allows the chatbot to access the web and pull data through third-party application plugins.
Multimodal AI systems, which combine conversational chatbots with systems that can produce synthetic media, such as highly convincing deepfakes, or include sensory abilities, such as seeing and hearingEuropol Report
ChatGPT and other similar language models are largely unsophisticated at the moment. However, they’re undergoing constant upgrades, with companies investing more and more in growing their capabilities.c
