The Senate has opened final voting on the Employee AI Training Consent Act, a workplace-focused artificial intelligence proposal sponsored by Mississippi Sen. Donna Lowenthal.

The vote remains open, and no outcome has been determined. With only a portion of senators having voted, the current count is trending toward an extrapolated result of 31 ayes and 14 nays, with no present votes, if the vote were to close at this stage. The Senate has 50 seats.

Lowenthal’s measure would require employers to obtain separate written consent before transferring worker-attributable material for artificial-intelligence development. It also bars retaliation against workers who refuse such consent and authorizes Federal Trade Commission penalties.

The bill places a specific workplace question at the center of the Senate’s broader debate over artificial intelligence: what control employees should retain when material tied to their work may be used in developing AI systems. Its consent requirement is directed at employer transfers of worker-attributable material, rather than AI policy in general.

Final voting follows recent Senate action on the proposal, which has drawn attention because it would establish federal requirements around employee consent and AI development. The measure’s progress now depends on the full Senate tally, not the early projection.

The early returns offer an initial indication of support but are not a final vote count. Senators who have not yet voted could still change the projected result, and the measure cannot be described as passed or defeated while voting remains open.

The vote comes as artificial intelligence continues to move into the foreground of policy debates involving workplaces, regulation and institutional accountability. But the legislation before the Senate is narrowly focused on consent, retaliation protections and FTC enforcement tied to the use of worker-attributable material in AI development.

Lowenthal, a Mississippi senator and former business executive, is the bill’s sponsor. Her proposal has reached the final-vote stage at a time when lawmakers are weighing how far federal policy should go in setting rules for AI-related practices.

For now, the central fact is procedural as much as political: the Senate is voting, the count is incomplete, and the projected 31-to-14 split is only an in-progress estimate based on votes cast so far.