First off, I believe AI is a very interesting tool, with great potential for both beneficial and harmful uses. I went over the dangers of AI in my final post, so here will be more about the current uses and benefits of AI.
First off, privacy and AI do not mesh all that well. Not only is it largely unknown whether the new(er) large language models (LLMs), such as ChatGPT, are being trained on people’s personal data, but who knows if they are collecting the prompts sent to them. A lot of data and information is hidden in the prompts people send in. It can show how people write or speak, OpenAI has made it possible to talk with ChatGPT, you can even show it pictures.
Needless to say, the ability for ChatGPT to analyze and store your voice, as well as potentially see you or collect images you send it, is a significant privacy risk. Especially since it is so difficult to see what the AI is collecting or being trained on. Without that information, it is hard to protect your data from the AI. Likewise, with how much information the AI needs to continue learning, especially as they train new models, it is highly unlikely that individual’s data is not being included.
As for security, AI is capable of automatically searching for cybersecurity threats such as hackers or viruses. This allows companies to respond to attacks much quicker, saving millions of dollars. Because of this, companies are investing more and more into AI security. It is costly not to, companies with no AI security lose $5.36 million on average, while companies with only marginal AI security drop that number to $4.04 million.
On the other hand, attackers can use AI to optimize or alter their methods of attack, making them more robust or easier. As mentioned before, there are both benefits and risks to AI’s implementation into society.