HomeAIAI Policy & SafetyBaroness Campbell Gives First AI-Voiced Speech in UK Parliament

Baroness Campbell Gives First AI-Voiced Speech in UK Parliament

For the first time in years, Baroness Campbell of Surbiton addressed the House of Lords in something close to her own voice on 13 October, and she used it to argue about NHS funding.

The crossbench peer, who has spinal muscular atrophy, spoke in the second reading of the government’s Health Bill using a recreation of her voice built with ElevenLabs software. Before the debate, the Government Chief Whip, Lord Kennedy of Southwark, told peers she would “address the House using an AI-generated voice. This is a first in a British Parliament,” according to Hansard. He asked the House to allow her some leeway on the advisory five-minute speaking time.

How the voice was made

Campbell’s voice has weakened over the years, and since 2012 she has relied on a personal assistant reading her words in the chamber, an arrangement the Lords Procedure Committee approved that year. She is the only peer allowed to bring an assistant into the chamber, according to Disability News Service.

She found ElevenLabs’ voice cloning tool herself, the company says. She uploaded old recordings of her speaking. Disability News Service says these came from around the time she began using a ventilator full-time in 2012, and included media interviews such as her appearance on BBC Radio 4’s Desert Island Discs that August. The software produced a voice matching the tone, rhythm and pitch of her old one. She then turned her written speech into audio with ElevenLabs’ text-to-speech tool, and it was played in the chamber through an assistive device. Disability News Service reported that the committee granted permission as a reasonable adjustment under the Equality Act.

“I used AI to speak my own words, not to put words into my mouth. The speech, the arguments and the views are entirely mine,” she said in a statement released by ElevenLabs. “It’s my voice, my style, my words.”

What she told the Lords

In the chamber she thanked the Procedure Committee, the technical department and Lords staff for “cutting-edge technology that will enable me to contribute to debates in my own, once-clear voice before it begins to deteriorate due to my disability”. Then, with some humour: “may I introduce noble Lords to Baroness Campbell mark 2?”

The substance was personal health budgets, the NHS funds people in England can use to arrange their own care. Campbell, who relies on one, welcomed the Bill but warned against the return of an “institutional medical model” of disability. She urged the government to keep direct payments, which let users control the money themselves, and raised the case of Lucinda Ritchie, who she said had been forced into residential care from hospital against her will. She ended: “I am not a medical diagnosis but a wife, grandparent and working member of the House of Lords.”

Health minister Baroness Merron told peers “we have a world-first today” and thanked the parliamentary authorities for “ensuring that accessibility prevails in this House”. The Lord Bishop of Newcastle noted “the historic nature of her excellent speech”.

An adjustment that keeps the words hers

The most important detail is the one Campbell stressed: the software read a speech she wrote. Nothing was generated except the sound. That separates this use from the voice clones that worry regulators and fraud teams, where a voice is copied to say things its owner never said. Here the owner consented, supplied the recordings and approved every word.

It also shows how Parliament chose to handle it. Rather than writing a rule about AI, the Lords treated the voice as an accessibility adjustment, the same route it used in 2012 for her assistant. That keeps the decision tied to one member’s need, with committee approval, instead of opening a general door to synthetic speech in the chamber. The Lord Speaker, Lord Forsyth of Drumlean, framed it the same way in comments released by ElevenLabs, acknowledging “justifiable concerns about some uses of AI” while backing this one.

Speech technology is moving quickly, from ElevenLabs’ real-time speech models to Mistral’s Voxtral 3 and Block’s Dessa voice model. A peer arguing about care budgets in her own recreated voice is one of the clearer cases of what that progress can be for.

The limits she named herself

Campbell was candid about the catch. She told Disability News Service that speeches must be written and recorded in advance, so the voice cannot help with unscripted interventions or replies. It also needs a good recording of the person’s voice from before it was lost or deteriorated. She had broadcast interviews to draw on; many people with progressive conditions will not, unless they record their voices early. In Spain, ElevenLabs and the patient association adELA now offer people with ALS free voice restoration, using new samples recorded while speech is still clear or old recordings.

ElevenLabs says its technology has been used to restore the voices of more than 11,000 people and that it aims to reach one million people living with voice loss. That target is the company’s own. For the House of Lords, the more immediate question is the next request: Campbell said she hopes other disabled people who struggle with public speaking will see that speech “does not have to be a barrier to inclusion at the highest level”.

Sources

Related stories

Follow @tsnmediahub on X for breaking AI news as it happens.

Follow on X

Share this story

Latest stories

  1. TensorWave and OpenAI Agree a 457MW AMD MI455X Deal
    AI Chips & Compute
  2. OpenAI Evaluator Pilots Start With Apollo and METR
    AI Policy & Safety

More in this category

Latest stories

  1. TensorWave and OpenAI Agree a 457MW AMD MI455X Deal
    AI Chips & Compute
  2. OpenAI Evaluator Pilots Start With Apollo and METR
    AI Policy & Safety

Explore the TSN Atlas: Every topic we cover, mapped. Click a bubble to dive in.