Many customers are captivated by Gen AI, employing new technologies for a variety of personal and corporate purposes. 

However, many people ignore the serious privacy implications.

Is Generative AI all sunshine and rainbows?

Consumer AI products, such as OpenAI's ChatGPT, Google's Gemini, Microsoft Copilot software, and the new Apple Intelligence, are widely available and growing. However, the programs have various privacy practices in terms of how they use and retain user data. In many circumstances, users are unaware of how their data is or may be utilized.

This is where being an informed consumer becomes critical. According to Jodi Daniels, chief executive and privacy expert of Red Clover Advisors, which advises businesses on privacy issues, the granularity of what you can regulate varies depending on the technology. Daniels explained that there is no uniform opt-out for all technologies.

Privacy concerns

The rise of AI technologies, and their incorporation into so much of what customers do on their personal computers and cellphones, makes these problems much more pressing. A few months ago, for example, Microsoft introduced its first Surface PCs with a dedicated Copilot button on the keyboard for rapid access to the chatbot, fulfilling a promise made several months previously. 

Apple, for its part, presented its AI vision last month, which centered around numerous smaller models that operate on the company's devices and chips. Company officials have spoken publicly about the significance of privacy, which can be an issue with AI models.

Here are many approaches for consumers to secure their privacy in the new era of generative AI.

1. Use opt-outs provided by OpenAI and Google

Each generation AI tool has its own privacy policy, which may include opt-out choices. Gemini, for example, lets customers choose a retention time and erase certain data, among other activity limits.

ChatGPT allows users to opt out of having their data used for model training. To do so, click the profile symbol in the bottom-left corner of the page and then pick Data Controls from the Settings header. They must then disable the feature labeled "Improve the model for everyone." According to a FAQ on OpenAI's website, if this is disabled, fresh talks will not be utilized to train ChatGPT's models.

2. Opt-in, but for good reasons

Companies are incorporating modern AI into personal and professional solutions, like as Microsoft Copilot. Opt-in only for valid reasons. Copilot for Microsoft 365, for example, integrates with Word, Excel, and PowerPoint to assist users with activities such as analytics, idea development, and organization.

Microsoft claims that it does not share consumer data with third parties without permission, nor does it utilize customer data to train Copilot or other AI features without consent. 

Users can, however, opt in if they like by logging into the Power Platform admin portal, selecting settings, and tenant settings, and enabling data sharing for Dynamics 365 Copilot and Power Platform Copilot AI Features. They facilitate data sharing and saving.

3. Gen AI search: Setting retention period

Consumers may not think much before seeking information using AI, treating it like a search engine to create information and ideas. However, looking for specific types of information utilizing gen AI might be intrusive to a person's privacy, hence there are best practices for using such tools. Hoffman-Andrews recommends setting a short retention period for the generation AI tool. 

And, if possible, erase chats once you've gathered the desired information. Companies still keep server logs, but they can assist lessen the chance of a third party gaining access to your account, he explained. It may also limit the likelihood of sensitive information becoming part of the model training. "It really depends on the privacy settings of the particular site."