UK and US Announce Joint Collaboration for AI Safety Testing

• The UK and US have signed a memorandum, announcing their collaboration to test the safety of AI models
• Two AI Safety Institutes have already been launched in both countries. While the one in the UK is up and running, the US institute will soon catch up
• The UK and the US are also trying to partner with other countries to expand their AI safety testing efforts

On Monday, the US and the UK signed an agreement announcing their partnership to test the most powerful AI models.

The agreement was signed by the U.S. Secretary of Commerce, Gina Raimondo, and Michelle Donelan, the U.K. Secretary of State for Science, Innovation, and Technology.

The two countries already have their own dedicated AI Safety Institutes that were inaugurated just one day apart, around the same time as the AI Safety Summit was held at Bletchley Park (last November).

For context, this is the summit where companies like OpenAI and Google DeepMind agreed to allow testing by third-party AI institutes before launching a new model in the market.

The collaboration is set to be effective immediately.

Speaking about the agreement, Raimondo admitted that it's no secret that AI is here to stay - it's the ultimate technology of this generation.

So instead of running away from the risks, they have decided to face them head-on and develop a framework that will allow both the institutes to address problems and threats that AI imposes on our society.

I think of the agreement as marking the next chapter in our journey on AI safety, working hand in glove with the United States government.Michelle Donelan

She also believes that while it's just the UK and the US collaborating now, they will soon become the driving force that brings together a network of institutes focused on AI safety testing.

The UK facility has been up and running for quite some time now. This is why Donelan believes this collaboration will also benefit the US as they'll be able to rely on the UK in the early stages of establishment and formalization.

Here's how the partnership will work:

• The two countries are planning to develop a common approach to AI safety. This means they'll be using the same testing methods and infrastructure.
• Also, to be in sync with each other, the countries will exchange employees and share any new information in accordance with the terms of their respective national laws and contracts.
• Last, but not least, they'll soon perform joint AI safety testing on a model available to the public.

Apart from their partnership, both countries are also trying to strike similar collaborations with other nations to expand their AI safety testing practices.

Other Steps Taken To Regulate AI

Starting from last year, both the US and UK have taken a series of steps to help the public benefit from AI without any risks.

• In February the UK government announced that they will be investing more than 100 million to prepare the country to safely handle AI. Part of that investment will also go into training and upskilling regulators around the nation.
• Similar steps have also been taken by the US which recently announced that every Federal agency using AI tools will need to put up a public list of tools they are using, reason for use, and a risk evaluation report of each.
• They have also been instructed to appoint a chief AI officer along with a governing body that will ensure that the agency is following all the new AI regulations.

Two developed nations showing keen interest in AI safety is a positive sign for the development of safe technology.

The post UK and US Announce Joint Collaboration for AI Safety Testing appeared first on The Tech Report.