Keep up with the innovative tech transforming business
Tech Brew keeps business leaders up-to-date on the latest innovations, automation advances, policy shifts, and more, so they can make informed decisions about tech.
One of the world’s strongest AI laws went into effect last week, with implications for companies in the EU and beyond.
In the works for years, the EU’s sweeping AI Act will enact new rules around the safety, sustainability, and bias of AI systems, especially those to be used for applications considered high-risk, like law enforcement, hiring, or critical infrastructure. The law took effect August 1, with some provisions to roll out over the next two years.
Like the EU’s General Data Protection Regulation (GDPR) before it, the law could have far-reaching effects for tech giants that do business in European markets, as well as set a benchmark for other jurisdictions.
Conformity check: According to a briefing issued by the European Parliament, companies selling AI products determined to be high-risk will have to carry out a conformity assessment before they can offer them in EU markets. That check encompasses “testing, data training, and cybersecurity” requirements, and could also be accompanied by a “fundamental rights impact assessment.”
A May report from the Brookings Institution predicted that AI companies operating internationally will likely choose to comply with those rules to avoid forgoing access to a lucrative market. They could also possibly develop region-specific models.
“Given the importance of the European market, international companies could be expected to align some of their AI governance practices with the AI Act to maintain access to the European Union’s internal market,” Brookings’ authors wrote.
Business response: When Microsoft inked a deal with French AI startup Mistral in February, Gartner Distinguished VP Analyst Jason Wong told us it could herald more geographically strategic deals that account for a new global regulatory landscape. Meanwhile, IBM partnered with German AI startup Aleph Alpha.
Meta recently said it would withhold its future multimodal Llama AI models in the EU, Axios reported last month, although the issue is reportedly related to GDPR restrictions on social posts used in training data rather than the AI Act.
Getting in line: A host of different international efforts meant to better align AI regulations around the world could also mean that the AI Act eventually serves as more of a blueprint for rules beyond EU borders. But while President Joe Biden signed an expansive executive order last fall that covers many of the same topics as the AI Act, efforts to pass more concrete federal legislation in the US have stagnated.