What the man with the red flag teaches us about AI regulation: David Richards
In 1865, before the motor car had arrived, Parliament decided that the self-propelled road vehicle was not to be trusted on its own. The Locomotives Act required a crew of three: a driver, a stoker and a man on foot, 60 yards ahead, warning the public that a machine was coming, with a red flag by day and a lantern by night.
The vehicle behind him could move at two miles an hour in town and four in the country. This was not wholly irrational. Horses bolted, roads were bad and steam engines were noisy.
But the effect was absurd. While Benz, Daimler and the American pioneers got on with making cars work, Britain made the new machine follow a pedestrian. The flag disappeared in 1878, but the escort survived until 1896. For 30 years, a country proud of its engineering slowed the future to walking pace.
I thought of that man this week while reading Sarah Breeden’s speech at Sintra. Breeden, the Bank of England’s deputy governor, was talking about autonomous AI agents in finance: systems that do not merely produce analysis, but act on it. In markets, she warned, they could amplify stress. Not because one model goes berserk, but because many models may do the same sensible-looking thing at once. Same data, same signal, same exit.


The Bank is looking at guardrails: circuit breakers and kill switches that could halt trading if models misfire. A trading system that acts at machine speed cannot be supervised like a junior dealer with a phone and a bad lunch.
Others are asking the same question. The Bank for International Settlements has warned that the AI investment boom could become a long bust. The Financial Stability Board is pressing for safeguards around autonomous agents. Existing rules may not be enough when the technology acts faster than those overseeing it.
That is a serious question. A kill switch on a runaway market is not the same as a red flag on a road. The trouble begins when regulators forget which one they are holding.
There is an old official habit of making the unfamiliar move slowly until important people feel less nervous. It looks prudent. Often it is only a lack of understanding dressed up as caution. The machine is new, so the state reaches for something old: a licence, a committee, a man with a flag.
History is littered with these panics. When margarine threatened butter, some American states forced it to be dyed an unappetising pink so that nobody mistook it for the real thing. In 1994, Westminster became frightened enough of rave culture to put “repetitive beats” into criminal law. Autechre answered with Flutter, a track arranged so that no bar repeated exactly.
And the fear is not only the state’s. In 2019 OpenAI declared its own GPT-2 model too dangerous to release. It had 1.5 billion parameters. The models that now help with homework, job applications and medical results are a thousand times larger. The feared monster became a useful tool.
The state rarely stops the new thing for long. It only makes sure it happens somewhere else first.
So yes, regulate AI in finance. Test the models, demand audit trails, establish liability. Make firms prove that an autonomous agent can be stopped, interrogated and overruled. But do not pretend that slowing a technology is the same as understanding it.
Britain still marks the end of the red flag. At the first Emancipation Run in 1896, a motorist tore one in half and drove for Brighton, as the cars have done every November since. It remembers how long the nation that built the industrial age made a man walk in front of the future.
AI needs rules. But a serious country must know the difference between a brake and a flag. The brake stops a crash. The flag stops the journey.
David Richards MBE is managing partner of Yorkshire AI Labs