Google’s new Gemini-powered Bard raises privacy eyebrows
It’s been about a year since OpenAI launched ChatGPT. Now, everyone’s in a race to make their AI tech bigger and better. This week, Google announced significant developments in its Bard chatbot and it’s coming in guns blazing.
The new version of Google Bard brings a wow factor but at a cost. Wondering how it stacks up against ChatGPT and how well (or not-so-well) it protects your data? You came to the right place.
New kid on the block
The significant difference between Google Bard and ChatGPT is how the AI is trained.
ChatGPT was initially trained on an undisclosed data set that only went to 2021. Today, the chatbot can access current information online, but only for users who pay for the Plus or Enterprise subscription.
On the flip side, Google Bard is powered by Gemini, which Google says is trained on an “infiniset” of data and has access to the internet in real time. It’s also free for all users, which is a huge advantage.
Gemini is also “natively multimodal,” meaning it’s trained on images, video and audio. ChatGPT is trained on text alone. This multimodal training makes Bard better at tasks like summarizing content, brainstorming, writing and planning.
Gemini AI comes in three flavors: Ultra, Pro and Nano. Bard is currently powered by Pro, which Google claims is best for a “wide range of tasks.”
Nano is best for on-device tasks. To be released in early 2024, Ultra is even more powerful and well-suited for highly complex tasks. We wouldn’t trust it to build a rocket, at least not yet.
The fine print
Here comes the most critical question: How does Bard protect your privacy?
When you ask the chatbot itself, it’ll tell you that Bard is better at protecting your information than ChatGPT because it doesn’t collect or store personal user information. In other words, Bard won’t track your browsing data or use your information for advertising purposes.
However, this is pretty misleading. You have to be logged into your Google account to use Bard and you can’t access it through an anonymous account. That doesn’t sound very … private.
Like ChatGPT, Bard allows users to delete their prompts but doesn’t necessarily delete conversations.
Google’s privacy policy says Bard conversations reviewed or annotated by human reviewers are not deleted when you delete your activity because they are kept separately and not connected to your Google Account. Instead, they’re retained for up to 3 years.
Even if you turn Bard activity off, your conversations are saved in your account for up to 72 hours. We’ll file that under things that make you go “hmmm.”
Technically, that means anything you share with Bard gets stored and maintained by Google. Where it goes from there is incredibly murky.
Our lips are sealed
If you’re down to give Bard a shot, proceed with caution. Assume everything you put into the chatbot will be saved and used to train Gemini to become an AI superpower.
Here are some things you shouldn’t share:
- Personally identifiable information (PII), including your full name, Social Security number or phone number.
- Private medical or mental health information.
- Passwords.
- Financial information, both personal and business-related.
- Business trade secrets or confidential work information.
- Sensitive life information, like relationship issues or lawsuits.
RELATED: These shopping apps collect and sell the most data.
Tags: chatbot, data, Google, Google Gemini, internet, privacy, security