What You May Have Missed #4
EU and US AI accountability / DeepMind's AlphaTensor improves algorithms / Bot translates videos into English / Self-Ask to improve GPT-3's Q&A perf / Boston Dynamics rejects robot weaponization
The EU and the US want to hold AI companies accountable
What happened
The EU unveiled last week a new bill on AI, called the AI Liability Directive, to empower consumers over AI companies. It complements the AI act and will become law in around two years, says Melissa Heikkilä for Tech Review. She writes:
“The new liability bill would give people and companies the right to sue for damages after being harmed by an AI system. The goal is to hold developers, producers, and users of the technologies accountable, and require them to explain how their AI systems were built and trained. Tech companies that fail to follow the rules risk EU-wide class actions.”
Tech companies argue that the new laws will hinder innovation while rights activists explain how it's insufficient: The burden of proof of harm lies on the consumer and the bill doesn't include indirect harms like unexpected failures.
President Biden announced on October 5th a similar approach to AI accountability. The AI Bill of Rights is similar to the EU's AI Act in intention but not in action. Indeed, it's more a blueprint for how AI accountability should work than a preface for regulation.
Here are the five principles, summarized by WIRED’s Khari Johnson:
“… people have a right to control how their data is used, to opt out of automated decisionmaking, to live free from ineffective or unsafe algorithms, to know when AI is making a decision about them, and to not be discriminated against by unfair algorithms.”
“The document is nonbinding and does not constitute US government policy,” says Heikkilä.
Why it’s relevant
The dark side of technology—AI is no different—is that it brings civilization forward while, at the same time, indifferently leaving behind the victims who suffer collateral harm.
The EU and the US's intentions to regulate AI companies, developers, and users and held them to account is a crucial first step to rebalancing the one-sided power dynamics, particularly harsh for already discriminated minorities.
If intention becomes law, they will soon have a voice and the possibility to defend themselves from those claiming to act in the name of progress.
Tech companies argue that strict regulation could have a “chilling” effect on progress, but that's not necessarily a bad thing if the alternative is stomping over those unlucky enough to fall at the tails of social privilege distributions.