Science · 1 sources covering this

How AI training models could compromise privacy

Credible studies have shown that querying an AI model with ‘anonymised’ data to extract personal details that it had memorised during training can be used to identify individuals and build their profiles

Share:WhatsApp X Reddit Facebook LinkedIn

Coverage spectrum

1 L · 0 C · 0 R

How 1 sources are covering this

The HinduLean L·58m ago
How AI training models could compromise privacy
Credible studies have shown that querying an AI model with ‘anonymised’ data to extract personal details that it had memorised during training can be used to identify individuals and build their profiles
Read at The Hindu →

Embed this bias spectrum on your site

Paste this iframe wherever you'd like the widget to appear. Resizes responsively; no JS required.

<iframe
  src="https://hindustansphere.com/embed/cluster/a64270f8d75ce0dfd6de4fe4379fe394e978868a"
  width="100%" height="320"
  style="border:0;max-width:680px"
  loading="lazy"
  title="How AI training models could compromise privacy"></iframe>

How AI training models could compromise privacy

How 1 sources are covering this

How AI training models could compromise privacy

More from Science

Firecrackers: Festivity or fallout?

Ebola reminder of urgent need to finish pandemic treaty: WHO chief

Vietnam scraps child organ donors proposal after backlash

Oocyte racket in Telangana: Women paid ₹20,000 as operators made around ₹3 lakh per cycle

Reader

Trust & legal

Across the network