AI Safety Summit: Who, What, Why?
By Fay Capstick
At the end of this week, the government will be holding a historic summit on AI safety at Bletchley Park. This blog will tell you everything you need to know about it.
What is going on?
Rishi Sunak, PM, has called for a summit to be held on the safety of AI. This will happen at the end of this week at Bletchley Park, famous for being home to the Enigma code breakers during World War II. Approximately 100 participants will be in attendance over the 2-day event.
Why?
As we have all been aware the capabilities of AI have grown hugely over recent months. This poses a potential problem for people, companies, academia, and countries. Without a uniform consensus of what we do with the genie now that it has started to leave the bottle, we could run into problems that threaten humanity.
Threats from AI could include the issue of deep fakes and their use to manipulate the political process, the loss of jobs, civil unrest, destabilisation of democracies, and the risk of triggering wars. These threats are all extremely serious. However, AI can also potentially bring about tremendous benefits for humanity and the planet. The key will be how we use AI, how others use AI, and whether regulations need to be implemented.
At this stage, the most important thing is for everyone to understand the risks and the scope of the threat that they pose. This is particularly important as private companies, and we assume the military, race to have the most powerful AI.
So what will it do?
The AI Safety Summit will look at these risks and how we can best approach and manage them. We only have one chance to get this right, and history will look back and judge how we approached the issue.
What is the problem?
One of the biggest hurdles is that the exact risk posed by AI will be hard to predict. All the different AIs being developed and how they are trained and might interact together make this harder.
Frontier AI?
This is referring to the forefront, the frontier, of AI development. It is also where some of the risks posed by AI are to be found. This summit will be questioning where the frontier of AI currently is and where it might be heading in the future.
What can this summit achieve?
The build-up period to the summit has had an opportunity for everyone to put forward their opinion on the debate. The actual summit will bring together countries, academics, civil society representatives, and companies, all of whom have already thought about the problem and want to bring their current conclusions to the table.
Does the summit have a set focus and objectives?
Yes, absolutely! The summit is going to be focusing on two types of risks posed by AI: misuse risks and loss of control risks.
Misuse risks: This could be where AI is used for bad purposes, be it by an individual, company, or country. This could be from cyber-attacks, biological or chemical releases, nuclear or military misuse, or damage to the systems that keep society running (infrastructure, banking, etc). All extremely worrying.
Loss of control risks: This is the idea, commonly found in sci-fi, that AI might develop a set of values and principles that do not align with humanity. This could be because an AI decides it has superior knowledge and reasoning. Issac Asimov addressed this with his Laws of Robotics, but obviously, these fail to help if AI has users with chaos-causing intent.
Can it really achieve anything, though?
Hopefully, if nothing else, it will make everyone aware of the problem and the need for everyone to engage to plot a way forward in the future. The government is saying that this will be the first AI Safety Summit and that the conversation will continue in future summits.
However, the goal would be for urgent action on AI safety to arise from this summit and to ‘promote best practice’. We can only hope that this is the case, as summits on global warming have repeatedly shown how words don’t lead to actions. The bigger worry might not be the AI but the users, as these will be almost impossible to control no matter how good the safeguards are.
Anything else?
Last week the government announced that the UK is establishing the world’s first AI Safety Institute. Its aim will be to ‘examine, evaluate and test new types of AI.’ The idea is that information will be shared so that there is a collaborative effort towards AI safety.
This all shows that the UK is taking a leading role in directing the conversation on AI safety and development. This is hopefully a positive step in ensuring that the UK will benefit from the boost to the economy and society that beneficial AI will bring, while managing any risk posed by frontier AI.
Final thoughts
At Parker Shaw, we have been at the forefront of the sector we serve, IT & Digital Recruitment and Consulting, for over 30 years. We can advise you on all your hiring needs.
If you are looking for your next job in the IT sector please check our Jobs Board for our current live vacancies athttps://parkershaw.co.uk/jobs-board