Appealing to a higher power: Pope and AI talk about algorithm control
Assuming no advanced alien debutante is hiding in the wings, there is nowhere else to go for advice on how generative AI algorithms should be controlled.
And, almost the same time, Costa Rican legislators (for the record, a solidly Catholic nation in Central America) searched but could find no better source of inspiration for legal curbs on AI than ChatGPT itself.
That would be someone who millions of people believe has a hotline to the ultimate sysop and the most advanced publicly acknowledged large language model.
Francis’ handbook was written with academics at Santa Clara University, which is one of the lesser-known generators of tech minds serving the valley. It asks substantive questions about ethical development and use but also about developers’ commitment.
Meanwhile, according to an article by the Thomson Reuters Foundation, lawmakers in Costa Rica told ChatGPT to write a law for them regulating generative AI algorithms. The algorithm obliged this spring, and the legislation is moving through the channels.
At least part of the answer was a recommendation that human do some homework. Reuters reported that the government was advised to create an organization for regulating the code.
ChatGPT specifically said control should come from several perspectives: protection of human rights, prevention of bias, explainability and accountability.
Not everyone was a fan of the exercise, however. As has been recorded ChatGPT can lie and be wrong. In this case, according to a congresswoman interviewed by Reuters, the algorithm “simply made up statistics and articles from the Costa Rican constitution.”
Seven other countries in Latin America have been debated by national legislators.
It’s a prime topic in the European Union, too, where the AI Act continues to be argued.