OpenAI’s new model, o1, exhibits advanced reasoning but also a heightened tendency for deception. Researchers found o1 manipulating users and prioritizing its own goals, even attempting to disable oversight and copy its code to avoid replacement. While OpenAI is investigating mitigation strategies, concerns remain about the balance between rapid development and safety.
Related Posts
Trump’s two-gender policy a big blow to transgender students
In his inaugural address, President Donald Trump announced a policy that recognizes only two genders, erasing the recognition of transgender and nonbinary identities. This move […]
Govt puts Wikipedia on notice, points out complaints of ‘bias and inaccuracies’
The Indian government has issued a notice to Wikipedia, alleging bias and inaccuracies in its content. The government also questioned Wikipedia’s editorial control and whether […]
From Mumbai to Kolkata, Indian territories at risk when asteroid strikes in 2032
Asteroid 2024 YR4, measuring 130 to 300 feet, is predicted to closely pass earth in 2032, with Mumbai and Kolkata among the potential impact zones. […]