News

Artificial Intelligence labs locked in ‘out of control race’, warn tech leaders

By on

AI experts and business leaders have petitioned AI labs to pause training of systems ‘more powerful than GPT-4’ until risks have been planned for and managed.


GPT-4 is the next iteration of the technology behind ChatGPT, a natural language processing tool driven by AI technology. It is designed to produce human-like text or dialogue and was released on 14 March.

In an open letter signed by leaders including Apple co-founder Steve Wozniak, Twitter CEO Elon Musk, and Yoshua Bengio, founder of AI research institute Mila, it warns that there must be checks on the rate of progress.

“Advanced AI could represent a profound change in the history of life on Earth, and should be planned for and managed with commensurate care and resources,” said the letter published by the Future of Life Institute, a non-profit for mitigating risks of transformative technology. “Unfortunately, this level of planning and management is not happening.”

HSE is expected to draw up guidance for organisations on implementing AI in a safe way. Photograph: iStock

It said that AI labs were locked in an “out-of-control race” to further the technology which even they may not “understand, predict, or reliably control”.

It says there are fundamental questions to ask before the technology is allowed to progress at the speed it is currently going: “Should we let machines flood our information channels with propaganda and untruth? Should we automate away all the jobs, including the fulfilling ones? Should we risk loss of control of our civilization?”

Meanwhile, HSE is expected to draw up guidance for organisations on implementing AI in a safe way according to five principles outlined in a government white paper.

These include safety and robustness, asking that applications of AI should ‘function in a secure, safe and robust way where risks are carefully managed.’

Fairness is another principle, i.e. that AI should be used in a way which complies with the UK’s existing laws, for example, the Equality Act or GDPR.

Legislation may be introduced next year to ensure consistent application of the guidance by regulators. “Given the pace at which AI technologies and risks emerge, and the scale of the opportunities at stake, we know that there is no time to waste if we are to strengthen the UK’s position as one of the best places in the world to start an AI company,” says the paper. “In collaboration with regulators, we are already exploring approaches to implementing the framework and will scale up this activity over the coming months.”

The government is also asking individuals and organisations to respond to its consultation, which runs until 21 June, found at the end of the white paper.

A pro-innovation approach to AI regulation: government's white paper and consultation is open until 21 June. Have your say here

NEWS


Wellbeing Conference 2024

Employers urged to listen to staff when designing workplace wellbeing strategies

By Kerry Reals on 26 April 2024

Conversation is king when it comes to implementing effective workplace wellbeing strategies, a panel of experts told a recent conference organised and hosted by British Safety Council.



ILO Climate Change Report Manal Azzi

Climate change could expose 70 per cent of workers to health risks, warns UN

By Kerry Reals on 22 April 2024

More than 70 per cent of the global workforce could be exposed to serious health hazards because of climate change, and existing occupational health and safety legislation may not be sufficient to protect workers, according to a new report by the International Labour Organization (ILO).



People Walking Past Job Centre iStock

Calls for wellbeing focus intensify as UK economic inactivity rate rises further

By Kerry Reals on 16 April 2024

The number of working-age people in the UK who are neither employed nor looking for a job rose again in the December to February period, prompting calls for more investment in training and a greater focus on the health and wellbeing of the nation.