Shane Legg launches institute for responsible AGI development and safety
Technology
Shane Legg, who helped start DeepMind, just launched a new institute focused on making sure advanced AI is developed responsibly.
With AI now writing code and handling tasks on its own, Legg wants better safety controls so things don't get out of hand.
The DeepMind Institute will spark conversations about how artificial general intelligence (AGI) could affect society, touching on everything from science and education to policy and philosophy.
Legg: 50-50 minimal AGI by 2028
Legg isn't convinced by claims that AGI is already here: he gives it a 50-50 shot by 2028 for even a "minimal" version.
He also said Amodei's call to slow but not pause the release of frontier models was "interesting directionally" and "worth considering."