European Union’s ambitious new A.I. regulations

Until now, artificial intelligence has been a kind of Wild West world with few rules governing the use of the new technology. Now, that's starting to change. The european Union proposed a sweeping new artificial intelligence act, the first of its kind that would ban some uses of the technology, for instance, using AI for ubiquitous public surveillance and scoring people based on their behavior, such as is done in china with the social credit system, there would be banned outright. Other uses of A. I would be governed according to their risk, considered high risk would be algorithms used in lending and insurance, healthcare, employment, those that determine access to education or training or where A I is used in law enforcement. In these cases, a government or business using the A. I. Would have to show that had undertaken a risk assessment and that it had put in place concrete steps to limit any potential harm In other areas, such as advertising. The use of Ai would be considered low risk and would not be subject to the same requirements. The penalty for violating the regulations large up to €20 million four of the company's global revenues, whichever is larger. This law will have potentially global impact. That's because any company with european employees or customers will be subject to it. Most U. S. Companies can't afford to abandon the EU as a market and training a separate AI system for different parts of the world is impractical. Plus most AI software performs better, the more data you feed it. So restricting an AI is training data to just one part of the world may mean it will perform less well than an AI system that's been trained on a larger, more global dataset. We could see this law in place by 2023. It's expected to take about two years for it to make its way through the ease legislative process. Meanwhile, in the U. S. The Federal Trade Commission has recently reiterated that it plans to crack down on any businesses that use A. I in a way that discriminates against protected class of people. It also plans to take action against those that are deceptive or manipulative in the way that they sell their AI systems. So while it's four may not be entirely clear yet, AI regulations coming and that is your time.