What risks do advanced AI models pose in the wrong hands?

News Snapshot:

The Biden administration is poised to open up a new front in its effort to safeguard U.S. AI from China and Russia with preliminary plans to place guardrails around the most advanced AI models, Reuters reported on Wednesday. Government and private sector researchers worry U.S. adversaries could use the models, which mine vast amounts of text and images to summarize information and generate content, to wage aggressive cyber attacks or even create potent biological weapons. Here are some threats posed by AI: Deepfakes and misinformation Deepfakes - realistic yet fabricated videos created by AI algorithms trained on copious online footage...