New York State passes RAISE Act for frontier AI models
TIL Creatives
Synopsis
New York lawmakers have approved the Responsible AI Safety and Education (RAISE) Act, mandating transparency and safety measures for frontier AI models. Supported by AI experts, the Act requires AI labs to release safety reports and report incidents, with penalties for non-compliance. India's AI adoption is growing, prompting increased demand for AI trust and safety professionals.
By ETtech
In a first of its kind, New York state lawmakers have passed the Responsible AI Safety and Education (RAISE) Act, to prevent frontier AI models by OpenAI, Google or Anthropic from contributing to disaster scenarios, including the death or injury of more than 100 people, or more than $1 billion in damages.
The legislation comes as a reform for the previous AI safety bill, which was ultimately vetoed. The AI safety bill targeted only large-scale models and didn’t address high-risk deployment or smaller but potentially dangerous models.
The Act is now awaiting the approval of New York governor Kathy Hochul, who can either sign it, send it back for amendments, or veto it.
The key provisions of the proposed RAISE Act include-
In a recent global survey by IBM, it was revealed that AI adoption in India is higher than in other countries. However, this is more experimentation while adoption at scale still lags.
Hiring in this space has surged 36% year-on-year, and the demand for AI trust and safety professionals is expected to grow by 2530% in 2025, data from Teamlease Digital showed.
ADVERTISEMENT
According to a TechCrunch report, the legislation has been supported by top AI experts Geoffrey Hinton and Yoshua Bengio. If converted into a law, it would be the first set of legally mandated transparency standards for frontier AI labs.The legislation comes as a reform for the previous AI safety bill, which was ultimately vetoed. The AI safety bill targeted only large-scale models and didn’t address high-risk deployment or smaller but potentially dangerous models.
The Act is now awaiting the approval of New York governor Kathy Hochul, who can either sign it, send it back for amendments, or veto it.
The key provisions of the proposed RAISE Act include-
- Requires AI labs to release safety and security reports on their frontier AI models
- In case of AI model behaviour or bad actors affecting the AI systems, the labs are mandated to report such safety incidents.
- Failure to comply with brings civil penalties up to $30 million
ADVERTISEMENT
Coming to similar perspectives on India’s AI ecosystem, analysts believe that while AI will be a catalyst for India’s economic growth, guardrails and governance will be key to adopt the technology safely and to build resilience amid possible disruption.In a recent global survey by IBM, it was revealed that AI adoption in India is higher than in other countries. However, this is more experimentation while adoption at scale still lags.
ADVERTISEMENT
To mitigate safety risks, tech multinationals and global capability centres (GCCs) in the country have started looking for specialised AI trust and safety roles, ET reported in March. Hiring in this space has surged 36% year-on-year, and the demand for AI trust and safety professionals is expected to grow by 2530% in 2025, data from Teamlease Digital showed.
