Adam Schiff says AI needs legislation, not another voluntary pledge

Senator Adam Schiff says Congress should legislate on AI sooner than it did on social media, arguing that voluntary promises from technology companies are not enough. In an interview with The Verge, he points to questions ranging from copyrighted training material to AI safety and use in the military.

Watch Desk analysis

What happened

Schiff, a California Democrat on Senate committees covering intellectual property, privacy, technology and antitrust, says lawmakers should first require companies to retain information about copyrighted material used to train their models. That, he argues, would give Congress a basis to debate what counts as fair use.

He also supports creating a new agency to regulate AI. Schiff says companies have publicly called for regulation while signing a non-binding White House safety pledge, and argues that the technology’s potential harms make delay a poor option. His remarks are in The Verge’s interview with Schiff.

Why it matters

The interview brings two practical questions into sharper focus: what training data AI companies should have to disclose or retain, and whether voluntary safety commitments can do the work of enforceable rules. Those choices affect creators, technology companies and anyone relying on AI systems.

Schiff is making a case for legislation, not announcing an agreed bill or settled congressional plan. His support for a new regulator is one proposal in a debate that also involves courts, existing agencies and the limits of government power.

Our read

The useful part of Schiff’s argument is its insistence that “regulate AI” is not one question. Training-data records, safety oversight and military use raise different issues and need more than a handshake and a hopeful press photo. The hard work is turning that urgency into rules that can survive scrutiny and do something measurable.

What to watch

• Whether Schiff or other senators introduce concrete proposals on training-data records or AI oversight.
• How lawmakers distinguish AI rules from restrictions on speech and other protected activity.
• Whether the White House pledge leads to enforceable commitments or remains voluntary.

Discussion spark: Should AI companies have to retain and disclose what copyrighted material they use for training, or would that put too much power over a complex legal question in government’s hands?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.