|
Top headlines
Lead story
If the marketing departments of tech companies are to be believed, humanity seems to be undergoing at least one to two groundbreaking revolutions per ad space. In reality, few technologies are as immediately earth-shattering as claimed. At best, a technology doesn’t live up to the hype but at least mostly works as advertised. At worst, it not only fails to deliver – it also harms its users.
Generative AI is a prime example of how a morass of claims of what a technology is capable of can muddle its actual capabilities. Many of the companies that bought into the hype about its ability to transform their workflows and bottom lines early on have instead seen these efforts fail. And casual chatbot users seeking to use AI to build relationships have found themselves instead breaking them.
If you can’t take what tech companies tell you at face value, how do you determine which claims are real and which are hyperbole?
One way might be through the scientific concept of validity – that is, how sound, trustworthy and dependable a claim is, measured from various angles and varying methods.
“Validity is the ultimate verdict of whether a scientific claim accurately reflects reality,” write Kai R. Larsen, Roman Lukyanenko and Thomas H. Davenport – of the University of Colorado Boulder, University of Virginia and Babson College, respectively. They developed a framework to help researchers test the validity of their findings across disciplines.
“For scientists, it’s a road map to ensure their inventions are rigorously evaluated,” the researchers write. “And for the public, it means knowing that the tools and systems they depend on – such as health apps, medications and financial platforms – are truly safe, effective and beneficial.”
[ Science from the scientists themselves. Sign up for our weekly science email newsletter. ]
|
