The context: more rules possible, plans unchanged
According to ETNews, the Democrats' victory in the U.S. midterm elections makes the risk of a tightening of AI rules more concrete. The newspaper cites the Wall Street Journal of November 9, according to which the technology chiefs of major U.S. companies do not, however, expect this to significantly affect their investments and AI adoption.
The article links this expectation to two factors: the models already available would be sufficient to meet corporate demand, and many companies would have already built their own security and governance systems. On this point, the material remains brief and does not quantify either investment plans or the actual uptake of the tools.
It is the source itself that flags that Democrats could introduce security checks before the release of new models. This is a possibility being envisaged, not a measure already approved: the text does not indicate timing or definitive content.
The companies' voices: rules as a trust factor
The article gathers the positions of several executives. Prakash Kota, chief transformation officer at UKG (workforce management software), argues that if rules offered certainty, even the most cautious companies would move toward AI: in his view, regulation could therefore accelerate adoption.
Dave Trett, CTO of Pearson, the education company, says his company already applies robust standards for governance, data protection, security, and human oversight, and believes that prudent regulation could increase trust and clarity. He adds that a slowdown in model development would not compromise AI use in the company, because the performance of current tools would still be largely untapped.
Doug Gilbert, CIO and CDO of Sutherland, an IT services company, is even more direct: according to him, day-to-day activities rarely require the most advanced models from OpenAI or Anthropic, and about 95% of the work can be done with existing models. These are statements by executives reported by the source, not independently verified data.
The new battleground: systems that govern models
The article reveals a shift in focus: companies' interest would be moving from the models themselves to the infrastructure that organizes and administers them, called an AI harness. According to the description provided, it is a software system that gives models memory and context on work data, allows different models to be replaced or combined, and makes it possible to control costs.
Alex Balazs, CTO of Intuit, a financial software company, says that a slowdown in model innovation would have a partial impact on companies, while today the company is concentrating substantial investments precisely on building an adequate model management system.
The issue of regulatory fragmentation remains open. According to Gilbert, it would be laws in states such as California, Colorado, New York, and Texas that set the concrete criteria for companies. Matt Calkins, CEO of Appian, a company active in AI automation, warns that differing rules from state to state would reduce consistency at the national level and considers federal intervention more appropriate on issues such as system alignment and intellectual property. Tony Bates, CEO and chairman of the board of Genesys, a customer experience technology company, shifts the problem to substance: the issue is having practical safeguards that strengthen trust without slowing innovation.
