As AI companies continue to claim that their systems could bring about the end of human civilization in the next several years—but should be permitted to continue development anyway, perhaps with some vague gesture at self-regulation or guardrails—the discussion on creating enforceable accountability rules around current harms seem to have been drowned out.
We at Mother Jones know that: Two years ago, we sued OpenAI and Microsoft for using our copyrighted work, along with vast amounts from others, to train their models. The documents uncovered as a result of the lawsuit show that executives and AI developers openly discussed the serious potential harms of their tech—even, in one case, acknowledging that they might have conducted “the largest theft of labor in human history”—yet carried on. As our CEO Monika Bauerlein laid out in an article published today,
A Microsoft executive warned that AI has started a “doom loop” that will “threaten the economic foundations of its essential suppliers” and eventually make the internet infinitely worse…Bottom line: We cannot leave these decisions to the AI executives. Not when it comes to stealing content, and not when it comes to the fate of humanity. Left to their own devices, they will make the wrong choices—even when they know those choices are bad for them, too.
—Alex Nguyen
No comments:
Post a Comment