Quiz: Can You Tell What’s Real and What’s AI?

Companies such as Adobe, Facebook owner Meta, Microsoft, Midjourney and OpenAI are pushing watermarks and other technological solutions to help folks identify artificially generated images. But at this stage, none is perfect.

Three out of four of these cute animals come from the imagination of an AI bot. The hamster is real.

Because an AI chatbot’s answers aren’t always reliable, using it to write a court document is not only unethical but also potentially damaging. In June 2023, a U.S. judge sanctioned New York lawyers for submitting a brief with fictitious legal citations that ChatGPT fabricated.

While the bad guys have a number of AI tools at their disposal, the government is fighting back. In February, the Federal Communications Commission banned robocalls that contain AI-generated voices and hopes to use AI itself to combat fraud. And President Biden issued an executive order in October to establish standards for AI safety, security and trustworthiness.

Prominent AI chatbots include Claude; Gemini, formerly Google Bard; and Copilot, formerly Bing Chat. All are rivals of OpenAI’s ChatGPT. Android is the smartphone operating system that Google designed but is not itself an AI bot.

Sora, not yet available to the public, can deliver video scenes of up to a minute based on a user’s text prompt. OpenAI is now testing Sora to prevent bias, misinformation and other potential dangers.