A Break to Restart Stronger
The open letter calls for a pause to allow AI laboratories and independent experts to develop and implement a set of shared security protocols for the design and development of advanced AI.
It is crucial for independent external experts to conduct rigorous audits and supervision of these protocols. This is to ensure that the systems are safe beyond all reasonable doubt.
For example, the implementation of detection systems to identify whether a text was produced by an AI or not. This will also involve the establishment of an ethics charter to avoid any drift. Finally, it will be necessary to put in place emergency mechanisms, in case AI becomes uncontrollable.
Indeed, no one can forget Microsoft’s attempts, whose AI connected to Twitter had started to make neo-Nazi remarks in just 24 hours.AI developers and regulators hold the authority to make researchers pause their work, and it is up to them to act now.
