The massive adoption of Chat GPT has caused problems for the cybersecurity area. Despite being able to increase global GDP by up to 7%, according to recent data from Goldman Sachs, new technology also poses risks to companies and organizations.
Even encouraging task automation, productivity and labor savings, the tool can assist in the creation and propagation of threats such as phishing and malware. The risk for companies would be linked to the high capacity to generate information based on Artificial Intelligence.
Highly creative, just a few questions asked by the chatbot, Potential criminals who have minimal technical knowledge can automate a network attack using codes launched by the tool. Thiago Marques, specialist in cybersecurity and partner at Add Value Security, explained how this happens.
As he, the NLP tool (Natural Language Processing), present no Chat GPT, performs a “learning from the past” function for the creation and generation of content. This mechanism can be easily used by criminals to create phishing campaigns and produce malicious Ransomware codes..
“With the version 4, the tool can deliver even more human results and process a greater amount of data, including links. Analyzing successful campaigns and existing data, create efficient phishing campaigns that can be carried out with little technical knowledge”, explains Marques.
Sensitive data exposes companies
In reality, the problem is exacerbated by the large amount of confidential corporate data available in Artificial Intelligence tools. Even Europol, European Union agency for police cooperation, recently claimed that the chatbot can be used for a series of frauds, cybercrime and disinformation.
These Europol analyzes were compiled in their first Tech Watch Flash report of the year, published now in March. The name of the survey was titled as: “GPT Chat – the impact of broad language models on law enforcement”.
OpenAI turns a blind eye
The technology is so controversial that OpenAI itself, Chat GPT developer company, has its own tools to detect texts generated by AI. This can help detect some content that is malicious. Other than that, phishing detection should be part of the overall network and infrastructure strategy.
“It is worth remembering that all these technologies are very recent. The benefits are enormous for everyone, just know the best way to use them. Cybersecurity professionals will not be replaced by AI or NPL, but they will need to learn how to best use the tools to continue defending themselves in the best possible ways.”, complete Marques.
