Microsoft President Brad Smith says that deep fakes, or content that looks real but is completely made up, are the biggest worry about artificial intelligence (AI). In a speech he gave in Washington, Smith talked about the need for good rules that make it easy to tell the difference between real photos or videos and ones made by AI, especially when they are used for bad things.
Smith talked about how important it is to fight deep fakes, bringing up concerns about cyber-influence operations from other countries. He talked about things that governments like Russia, China, and Iran have already done.
Smith said, “We’re going to have to deal with the issues surrounding deep fakes. We will have to deal with the kinds of things that the Russian government, the Chinese, and the Iranians are already doing that make us most worried about foreign cyber influence operations. He also said, “We need to take steps to protect against the use of AI to change legitimate content in order to trick or scam people.”
The President of Microsoft also pushed for licensing important types of AI and putting in place obligations to protect physical security, cybersecurity, and national security. Smith also called for updated export controls to keep AI models from being stolen or used in ways that would break a country’s export control rules.
Smith suggested that developers of powerful AI models use a “Know Your Customer”-style system to make things more clear and make it easier to spot videos that have been tampered with. A system like this would make it possible to track how AI is used and help people figure out where content made by AI came from.
Last week, Sam Altman, the CEO of OpenAI, talked to a Senate panel about his worries about how AI could affect the integrity of elections and how important it is to have rules. Altman also called for the world to work together on AI and for safety rules to be made easier to follow. But Altman said earlier this week that OpenAI might have to stop working if the EU decides to regulate AI too much. Top executives from the EU have also met with Sundar Pichai, the CEO of Google, to talk about how to regulate AI while the new policy is being made.