As AI Moves at Warp Speed New Tools May Help Prevent Mistakes

Most people know about OpenAI’s ChatGPT, funded, in part, by a massive infusion of cash from Microsoft. People also are aware of the privacy concerns of GPT-type tools.

In many cases, the damage is not intentional, but accidental. People submitting sensitive queries without thinking about it.

Also, as more and more software incorporates GPT-tools under the covers, the problem is only going to get worse.

But since capitalism always looks for opportunities to make money, new tools and techniques will appear. In this case, there may be so much money to be made that the tools may show up quickly. This likely will be good for us.

One new tool is PrivateGPT – a take on OpenAI’s name, I suspect.

PrivateGPT integrates with OpenAI’s platform and automatically redacts more than 50 types of personally identifiable data. Is it going to be perfect? Of course not. Will it help? Likely.

PrivateGPT sits in between you and that great Internet void and strips out health data, credit card information, dates of birth and a lot of other information.

Even OpenAI has warned users to be careful. Please don’t share any sensitive information in your conversations, says OpenAI’s user guide.

Samsung, for example, seems to have once again banned OpenAI’s tool after three different security incidents leaked equipment defects, buggy source code and minutes of an internal meeting.

This is going to be a challenge for companies, but the smart companies are going to meet this head on because AI is not going away. This includes contract issues, training and monitoring, among other things.

If you need help with this, please contact us.

Credit: Dark Reading

by