AI companies’ pledge: Activating a "kill switch" if necessary
AI companies OpenAI and Anthropic have promised to halt or pause their AI models if British intelligence agencies present evidence that they pose a security risk. This would involve the use of a so-called "kill switch," reports Sky News.
"If any authority came forward with demonstrable evidence that the model was unsafe or dangerous, we would naturally withdraw it from the market," said Liam Booth Smith, a representative for Anthropic, during a parliamentary hearing.
At the same time, they state that this can only happen if they retain control over the models.
Fund investments
Strategist: Pension savers should be aware of AI-related risks
There are many risks associated with AI development that Swedish pension funds should be aware of. This is according to Sasja Beslik, a strategist at the venture capital firm SDG, writing in an opinion piece for *DN Debatt*.
Based on this analysis, a number of measures should be taken, such as disclosing the extent of AI exposure held by funds like AP7 and imposing requirements on companies building data centers in Sweden.
"Capital without transparency is not asset management—it is speculation. Swedish savers have the right to know what their money is funding," writes Beslik.
Inga kommentarer:
Skicka en kommentar