Nevertheless, the warning about the risks of GPT-4 is already justified today. Because AI does not only harbor dangers when it becomes as clever as Albert Einstein. She already has superhuman abilities in niches that can be abused. It can take phenomena such as disinformation, spam and the influencing of political elections to a whole new technical level – through tailor-made propaganda texts or deceptively real fake pictures, such as the ones shown in the past few days of the alleged arrest of Donald Trump.
Also read: Can ChatGPT explain the economy?
What to do? AI developers should work with policymakers to dramatically accelerate the development of robust AI governance systems, the letter’s authors urge. And propose monitoring and tracking high-performance AI systems, introducing provenance and watermarking systems that help distinguish real from synthetic data, clarifying liability for damage caused by AI, and introducing well-equipped institutions to deal with the dramatic economic and political disruptions that AI will cause.