359
Productivity & Workflow355
Software Development249
Automation & Workflow238
Marketing & Growth207
AI Infrastructure & MLOps197
Photography & Imaging169
Writing & Content Creation195
Design & Creative172
Data & Analytics165
Customer Support157
Education & Learning144
Sales & Outreach133
Voice & Speech128
Operations & Admin119
Anthropic says its AI progress is speeding up, so it activated extra safety steps in its policy, days before an IPO that could value it at $1.5tn.
In short: Anthropic says its AI is improving fast enough that it has turned on extra safety controls, just days before its IPO.
Anthropic, the company behind the Claude chatbot, said it has activated part of a safety plan it created in 2023. The plan is called its “responsible scaling policy”, and it is meant to slow down or add checks if its AI starts advancing too quickly.
Anthropic disclosed this in an update to its IPO paperwork. An IPO is when a private company starts selling shares to the public on a stock exchange. The Financial Times reported the IPO could value Anthropic at about $1.5 trillion.
The trigger for these extra controls was Anthropic’s “overall rate of progress”. The company’s policy says it should add more oversight when it can “compress two years of AI progress into a single year”, based on how fast AI improved between 2018 and 2024.
A separate report from METR, a group that tests and compares AI systems, said Anthropic had made roughly two years of historical progress in 12 months. The report said a big reason was that Anthropic’s researchers are using AI tools to help build the next versions of their own AI, like using a power tool to build a better power tool.
Some AI researchers and safety campaigners criticized Anthropic, saying it is using safety messaging while still racing to develop more powerful systems.
If AI starts helping to build better AI with less human direction, progress could speed up in ways that are hard to manage. Anthropic itself warned in its IPO prospectus that its models could pose “catastrophic or existential risks” if misused or not properly secured.
Source: Financial Times