New Bill Holds AI Companies Responsible For Unsafe Use
A new AI bill in California looks to regulate the new technology, giving companies in the state legal requirements and liability. If passed, the law would hold the companies who develop the technology legally liable for harm done using their products and set up fail-safes to prevent major catastrophes. All eyes are on the divisive bill, which some view as much-needed regulation while others see it as a hurdle for innovation.
Opening Companies Up To Potential Lawsuits
Senate Bill 1047 is an attempt to regulate AI, creating clear legal liability when it comes to dealing with emerging technology. It would hold companies responsible in civil courts, but not criminal courts, for any models used in harmful ways.
The idea is that the threat of lawsuits would incentivize companies to prioritize product safety and fail-safes that prevent the misuse of their models.
In addition to opening companies to lawsuits from the public, the AI bill would allow the Attorney General to take legal action against companies.
This power would be reserved for cases in which a model either does significant harm to residents of California or poses a threat to those citizens. In these cases, the company would be considered negligent.
Safety Guidelines
In addition to holding companies responsible for their products, the AI bill would give developers working on the most powerful versions of AI new safety guidelines. The requirements include mandatory kill switches for all models, protections for whistleblowers, regular safety audits, and meeting certain cyber security requirements. These requirements would only apply to models that cost over $100 million to train or surpass a large computing power requirement.
A Vote Is On The Way
The Assembly Appropriations Committee has approved the AI bill but still has a few hurdles to jump through before it becomes law. The next step will be a vote on the full Assembly floor, which has to happen before August 31 giving it a limited window before it expires. Then it must be signed by Governor Gavin Newsom.
Unsurprisingly, the AI bill has become a source of controversy both in and out of the tech community. Many have called for stricter regulations around the development of AI models due to their potentially dangerous capabilities.
However, critics have accused the legislation as an unnecessary precaution that will significantly hamper the development process and weaken California’s economy.
Preventing Breakthroughs?
Due to the state’s position as the central hub for AI development, the California bill has potentially global implications. With the new technology playing a central role in misinformation campaigns and personal attacks, new regulations in California could curb the problems it is already creating.
Of course, regulation could also prevent breakthroughs in the emerging industry, which has the potential to transform numerous fields around the world.
High Stakes
If passed, Senate Bill 1047 could reshape how AI is developed by putting much more rigorous safety standards and legal obligations on the companies developing new models.
With the potential capability to devastate infrastructure, engage in cyberattacks, and even create bioweapons, the stakes couldn’t be higher. Of course, with so much anxiety and interest around the technology, more legislation at the national and even international level could be coming.
